recommender-systems/05_multimodality.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "05_multimodality.ipynb",
      "provenance": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU",
    "widgets": {
      "application/vnd.jupyter.widget-state+json": {
        "0e5f0348d4bb4d28a3935b7b28eec381": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HBoxModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HBoxModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HBoxView",
            "box_style": "",
            "children": [
              "IPY_MODEL_72b14c0707384ef4b2046c8221f0cb20",
              "IPY_MODEL_5c7270734d9b467dac9ee395de866215",
              "IPY_MODEL_b359be26e84b474cac45f72e54db3409"
            ],
            "layout": "IPY_MODEL_e831e2e7f6c341daa3177f5b4770b7ab"
          }
        },
        "72b14c0707384ef4b2046c8221f0cb20": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_189088aae00d47dd8b581616413c9373",
            "placeholder": "​",
            "style": "IPY_MODEL_c34ad3c74be044db9cf185eceed49625",
            "value": ""
          }
        },
        "5c7270734d9b467dac9ee395de866215": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "FloatProgressModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "ProgressView",
            "bar_style": "success",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_285374eb26094d3498828c85241e53ca",
            "max": 1,
            "min": 0,
            "orientation": "horizontal",
            "style": "IPY_MODEL_a425385ac6b941338a512b9a824a3a91",
            "value": 1
          }
        },
        "b359be26e84b474cac45f72e54db3409": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_c7045c02d9684a2bbf8abb90937fbda6",
            "placeholder": "​",
            "style": "IPY_MODEL_ff9da1a29d624420850f4f12fadf7e88",
            "value": " 106k/? [00:00&lt;00:00, 35.0kB/s]"
          }
        },
        "e831e2e7f6c341daa3177f5b4770b7ab": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "189088aae00d47dd8b581616413c9373": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "c34ad3c74be044db9cf185eceed49625": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "285374eb26094d3498828c85241e53ca": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": "20px"
          }
        },
        "a425385ac6b941338a512b9a824a3a91": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "ProgressStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "bar_color": null,
            "description_width": ""
          }
        },
        "c7045c02d9684a2bbf8abb90937fbda6": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "ff9da1a29d624420850f4f12fadf7e88": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "1f50664225f54fbdb5e900ef3279f97b": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HBoxModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HBoxModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HBoxView",
            "box_style": "",
            "children": [
              "IPY_MODEL_f310aba5515d4b1eb74f3a0e422f2644",
              "IPY_MODEL_1f13fb9c0979425aa03cfc9148f42dac",
              "IPY_MODEL_a4052a5171e441dbb168fbacf865fb32"
            ],
            "layout": "IPY_MODEL_f37087d5bf674d02ae75bcb5b75136d3"
          }
        },
        "f310aba5515d4b1eb74f3a0e422f2644": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_3c835cfb9c8346df9a7f29784fd0c390",
            "placeholder": "​",
            "style": "IPY_MODEL_69a532260693427b8985818cc9019203",
            "value": ""
          }
        },
        "1f13fb9c0979425aa03cfc9148f42dac": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "FloatProgressModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "ProgressView",
            "bar_style": "success",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_93819f2f8a1847f4911973d70b24248e",
            "max": 1,
            "min": 0,
            "orientation": "horizontal",
            "style": "IPY_MODEL_21540e10e9ae4054bcfdf5a0c5820070",
            "value": 1
          }
        },
        "a4052a5171e441dbb168fbacf865fb32": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_f73928de4e764904a75e25f8fd4a4ac1",
            "placeholder": "​",
            "style": "IPY_MODEL_199cf847db5d4d219a383abc28a8c6b3",
            "value": " 164k/? [00:00&lt;00:00, 68.7kB/s]"
          }
        },
        "f37087d5bf674d02ae75bcb5b75136d3": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "3c835cfb9c8346df9a7f29784fd0c390": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "69a532260693427b8985818cc9019203": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "93819f2f8a1847f4911973d70b24248e": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": "20px"
          }
        },
        "21540e10e9ae4054bcfdf5a0c5820070": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "ProgressStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "bar_color": null,
            "description_width": ""
          }
        },
        "f73928de4e764904a75e25f8fd4a4ac1": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "199cf847db5d4d219a383abc28a8c6b3": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "e10a12f760104a1ab3e181b33ec878bb": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HBoxModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HBoxModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HBoxView",
            "box_style": "",
            "children": [
              "IPY_MODEL_342c1ec0286e496e87874f41102521d7",
              "IPY_MODEL_faa730b609f84b60a58f07ac35290ebe",
              "IPY_MODEL_39de41ba3b1645c3b4c748effa0f03af"
            ],
            "layout": "IPY_MODEL_f2c786013d8f48b581d67f4e2edd898f"
          }
        },
        "342c1ec0286e496e87874f41102521d7": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_1e8f16db87ff4a27945cc6fa9d053296",
            "placeholder": "​",
            "style": "IPY_MODEL_1458df38c7f64297a88916fa52d5b960",
            "value": ""
          }
        },
        "faa730b609f84b60a58f07ac35290ebe": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "FloatProgressModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "ProgressView",
            "bar_style": "success",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_6cf235cfcea7430aadeaaf7e53c16de3",
            "max": 1,
            "min": 0,
            "orientation": "horizontal",
            "style": "IPY_MODEL_847ca2b8872c40cc886bded5e076388b",
            "value": 1
          }
        },
        "39de41ba3b1645c3b4c748effa0f03af": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_910379254e764589ba89499a00902aed",
            "placeholder": "​",
            "style": "IPY_MODEL_a95b99387abf4d77bf51ce07f80b9d02",
            "value": " 475k/? [00:00&lt;00:00, 5.46MB/s]"
          }
        },
        "f2c786013d8f48b581d67f4e2edd898f": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "1e8f16db87ff4a27945cc6fa9d053296": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "1458df38c7f64297a88916fa52d5b960": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "6cf235cfcea7430aadeaaf7e53c16de3": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": "20px"
          }
        },
        "847ca2b8872c40cc886bded5e076388b": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "ProgressStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "bar_color": null,
            "description_width": ""
          }
        },
        "910379254e764589ba89499a00902aed": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "a95b99387abf4d77bf51ce07f80b9d02": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "04c0fda0b1b14d85973e24befbd0997b": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HBoxModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HBoxModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HBoxView",
            "box_style": "",
            "children": [
              "IPY_MODEL_fe5af89a90b6460aa992b62938552d7b",
              "IPY_MODEL_7efbd2a55a7244a89c007269e067ec9a",
              "IPY_MODEL_c505ce61346d4dfb9c214322ac955b56"
            ],
            "layout": "IPY_MODEL_6b7afb98961a4a918ddb83879409046d"
          }
        },
        "fe5af89a90b6460aa992b62938552d7b": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_f7ad5692ade94c87be902a84acff243d",
            "placeholder": "​",
            "style": "IPY_MODEL_8156dc94e21940be9a556b06e5f818f8",
            "value": ""
          }
        },
        "7efbd2a55a7244a89c007269e067ec9a": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "FloatProgressModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "ProgressView",
            "bar_style": "success",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_18d7c30049f048f8a27bda6bf3a3b220",
            "max": 1,
            "min": 0,
            "orientation": "horizontal",
            "style": "IPY_MODEL_7aeb2766deb44ff1a4b04930f5b2c7e1",
            "value": 1
          }
        },
        "c505ce61346d4dfb9c214322ac955b56": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_77ae7062d88f4e5db5cd2b4fd917825c",
            "placeholder": "​",
            "style": "IPY_MODEL_3102c018ee354b2ba0be7389bca3519e",
            "value": " 14.9M/? [00:00&lt;00:00, 11.5MB/s]"
          }
        },
        "6b7afb98961a4a918ddb83879409046d": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "f7ad5692ade94c87be902a84acff243d": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "8156dc94e21940be9a556b06e5f818f8": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "18d7c30049f048f8a27bda6bf3a3b220": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": "20px"
          }
        },
        "7aeb2766deb44ff1a4b04930f5b2c7e1": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "ProgressStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "bar_color": null,
            "description_width": ""
          }
        },
        "77ae7062d88f4e5db5cd2b4fd917825c": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "3102c018ee354b2ba0be7389bca3519e": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "640728869a334399b926a0c1e3236c81": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HBoxModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HBoxModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HBoxView",
            "box_style": "",
            "children": [
              "IPY_MODEL_94c1106c817f41368c4e42d5455254d4",
              "IPY_MODEL_9b5f169748fd40a1b92502afe2592ac3",
              "IPY_MODEL_588760830bfb45d9b76b64ae9370988b"
            ],
            "layout": "IPY_MODEL_f04e7f82eb03434a9a6755197cb906a7"
          }
        },
        "94c1106c817f41368c4e42d5455254d4": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_aac8895239cc4100854c16b36ca89f59",
            "placeholder": "​",
            "style": "IPY_MODEL_292153a6e05d47d49cc901141aa7e347",
            "value": ""
          }
        },
        "9b5f169748fd40a1b92502afe2592ac3": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "FloatProgressModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "ProgressView",
            "bar_style": "success",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_9e531d3678be4fa38f6467d562ac5286",
            "max": 1,
            "min": 0,
            "orientation": "horizontal",
            "style": "IPY_MODEL_3d7e45fb56914571b447bd28f04ec470",
            "value": 1
          }
        },
        "588760830bfb45d9b76b64ae9370988b": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_7c94da4bbef3481aac2d28bbd165154f",
            "placeholder": "​",
            "style": "IPY_MODEL_41ed6a1c3b0a49ee9ab22a91f8659797",
            "value": " 32.8k/? [00:00&lt;00:00, 378kB/s]"
          }
        },
        "f04e7f82eb03434a9a6755197cb906a7": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "aac8895239cc4100854c16b36ca89f59": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "292153a6e05d47d49cc901141aa7e347": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "9e531d3678be4fa38f6467d562ac5286": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": "20px"
          }
        },
        "3d7e45fb56914571b447bd28f04ec470": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "ProgressStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "bar_color": null,
            "description_width": ""
          }
        },
        "7c94da4bbef3481aac2d28bbd165154f": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "41ed6a1c3b0a49ee9ab22a91f8659797": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        }
      }
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "xLXWx4uhqtYJ"
      },
      "source": [
        "<table class=\"tfo-notebook-buttons\" align=\"left\">\n",
        "  <td>\n",
        "    <a target=\"_blank\" href=\"https://colab.research.google.com/github/PreferredAI/tutorials/blob/master/recommender-systems/05_multimodality.ipynb\"><img src=\"https://www.tensorflow.org/images/colab_logo_32px.png\" />Run in Google Colab</a>\n",
        "  </td>\n",
        "  <td>\n",
        "    <a target=\"_blank\" href=\"https://github.com/PreferredAI/tutorials/blob/master/recommender-systems/05_multimodality.ipynb\"><img src=\"https://www.tensorflow.org/images/GitHub-Mark-32px.png\" />View source on GitHub</a>\n",
        "  </td>\n",
        "</table>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "wzPrmZOMtXt1"
      },
      "source": [
        "# Multimodality"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Bh3gYub8f6qZ"
      },
      "source": [
        "While preference data in the form of user-item interactions are the backbone of many recommender systems, such data tends to be sparse in nature. One way to address this sparsity is to look beyond the interaction data to the additional information associated with users or with items. The intuition is that items with similarity in \"content profiles\" would be correlated with similarity in preferences. Multimodality deals with how to model both preference data (one modality) and some content data either on user or item side (other modalities). In this tutorial, we see three forms of additional modalities, namely graph, text, and image, and investigate whether they add value to the the resulting recommendations."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "MqktDa7H2hKz"
      },
      "source": [
        "## 1. Setup"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "41oWCMUG2eC_",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "b2d78720-16cd-45a7-b5e7-f73e236d25e1"
      },
      "source": [
        "!pip install --quiet cornac==1.15.4"
      ],
      "execution_count": 1,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m18.5/18.5 MB\u001b[0m \u001b[31m48.2 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25h"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "YqKrDcGH2k7E",
        "outputId": "eb62313e-1229-4565-88b5-71a63f9624bb",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "source": [
        "import os\n",
        "import sys\n",
        "from collections import defaultdict\n",
        "\n",
        "import numpy as np\n",
        "import pandas as pd; pd.set_option(\"max_colwidth\", 0);\n",
        "%matplotlib inline\n",
        "import matplotlib.pyplot as plt\n",
        "import seaborn as sns\n",
        "\n",
        "import cornac\n",
        "from cornac.utils import cache\n",
        "from cornac.datasets import filmtrust, amazon_clothing\n",
        "from cornac.eval_methods import RatioSplit\n",
        "from cornac.models import PMF, SoRec, WMF, CTR, BPR, VBPR\n",
        "from cornac.data import GraphModality, TextModality, ImageModality\n",
        "from cornac.data.text import BaseTokenizer\n",
        "\n",
        "import tensorflow as tf\n",
        "\n",
        "print(f\"System version: {sys.version}\")\n",
        "print(f\"Cornac version: {cornac.__version__}\")\n",
        "print(f\"Tensorflow version: {tf.__version__}\")\n",
        "\n",
        "SEED = 42\n",
        "VERBOSE = False"
      ],
      "execution_count": 2,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "System version: 3.10.11 (main, Apr  5 2023, 14:15:10) [GCC 9.4.0]\n",
            "Cornac version: 1.15.4\n",
            "Tensorflow version: 2.12.0\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "r76coF3ZyWxb"
      },
      "source": [
        "## 2. Graph Modality\n",
        "\n",
        "In recommender systems, graph can be used to represent user social network or item contexts (e.g., co-views, co-purchases).  In this tutorial, we take the former as an example and discuss SoRec [2], a representative model for this class of algorithms. \n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "0GePqE99j5nr"
      },
      "source": [
        "### Social Recommendation (SoRec)\n",
        "\n",
        "SoRec model is based on matrix factorization framework. The idea is fusing user-item rating matrix with the user’s social network.  In summary, the *user-item rating matrix* ($R$) and the *user-user graph adjacency matrix* ($G$) are factorized with shared users' latent factors.  The user latent vectors in $\\mathbf{U}$ are shared to capture both user preferences as well as their social connections.  The rating prediction is obtained as $\\hat{r}_{i,j} = \\mathbf{u}_i^T \\mathbf{v}_j$, similarly to PMF model.\n",
        "\n",
        "To learn the model parameters, we minimize the following loss function:\n",
        "\n",
        "$$ \\mathcal{L}(\\mathbf{U,V,Z}|\\lambda,\\lambda_C) = \\frac{1}{2} \\sum_{r_{i,j} \\in \\mathcal{R}} (r_{i,j} - \\mathbf{u}_i^T \\mathbf{v}_j)^2 + \\frac{\\lambda_C}{2} \\sum_{g_{i,h} \\in \\mathcal{G}} (g_{i,h} - \\mathbf{u}_i^T \\mathbf{z}_h)^2 + \\frac{\\lambda}{2} \\sum_{i=1}^{N} ||\\mathbf{u}_i||^2 + \\frac{\\lambda}{2} \\sum_{j=1}^{M} ||\\mathbf{v}_j||^2 + \\frac{\\lambda}{2} \\sum_{h=1}^{N} ||\\mathbf{z}_h||^2 $$\n",
        "\n",
        "where $\\lambda_C$ is the relative importance of the social network factorization and $\\lambda$ is the regularization weight. \n",
        "\n",
        "Let's do a comparison between SoRec and its base model PMF on [FilmTrust dataset](http://konect.cc/networks/librec-filmtrust-trust/)."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "p2iyzuFFs1Fq",
        "outputId": "366696ab-e076-4a4a-cd26-060453cd9d6c",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 299,
          "referenced_widgets": [
            "0e5f0348d4bb4d28a3935b7b28eec381",
            "72b14c0707384ef4b2046c8221f0cb20",
            "5c7270734d9b467dac9ee395de866215",
            "b359be26e84b474cac45f72e54db3409",
            "e831e2e7f6c341daa3177f5b4770b7ab",
            "189088aae00d47dd8b581616413c9373",
            "c34ad3c74be044db9cf185eceed49625",
            "285374eb26094d3498828c85241e53ca",
            "a425385ac6b941338a512b9a824a3a91",
            "c7045c02d9684a2bbf8abb90937fbda6",
            "ff9da1a29d624420850f4f12fadf7e88"
          ]
        }
      },
      "source": [
        "K = 20\n",
        "sorec = SoRec(k=K, max_iter=50, learning_rate=0.001, verbose=VERBOSE, seed=SEED)\n",
        "pmf = PMF(k=K, max_iter=50, learning_rate=0.001, lambda_reg=0.01, verbose=VERBOSE, seed=SEED)\n",
        "\n",
        "ratings = filmtrust.load_feedback()\n",
        "trust = filmtrust.load_trust()\n",
        "\n",
        "user_graph_modality = GraphModality(data=trust)\n",
        "\n",
        "ratio_split = RatioSplit(\n",
        "    data=ratings,\n",
        "    test_size=0.2,\n",
        "    rating_threshold=2.5,\n",
        "    exclude_unknowns=True,\n",
        "    user_graph=user_graph_modality,\n",
        "    verbose=VERBOSE,\n",
        "    seed=SEED,\n",
        ")\n",
        "\n",
        "mae = cornac.metrics.MAE()\n",
        "\n",
        "cornac.Experiment(eval_method=ratio_split, models=[sorec, pmf], metrics=[mae]).run()"
      ],
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Data from https://static.preferred.ai/cornac/datasets/filmtrust/filmtrust.zip\n",
            "will be cached into /root/.cornac/filmtrust/ratings.txt\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "0.00B [00:00, ?B/s]"
            ],
            "application/vnd.jupyter.widget-view+json": {
              "version_major": 2,
              "version_minor": 0,
              "model_id": "0e5f0348d4bb4d28a3935b7b28eec381"
            }
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Unzipping ...\n",
            "File cached!\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.10/dist-packages/cornac/data/dataset.py:361: UserWarning: 3 duplicated observations are removed!\n",
            "  warnings.warn(\"%d duplicated observations are removed!\" % dup_count)\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            "TEST:\n",
            "...\n",
            "      |    MAE | Train (s) | Test (s)\n",
            "----- + ------ + --------- + --------\n",
            "SoRec | 0.6644 |    1.2729 |   0.3798\n",
            "PMF   | 0.7018 |    0.7729 |   0.2736\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "2oQ7J1_r2CGT"
      },
      "source": [
        "From the experiment, we see that SoRec achieves lower (better) MAE score as compared to PMF.  This improvement should be explained by useful information from user social network captured inside the model predictions.  Let's take a look at user factors and see if there are connections to the social network.  "
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "QUFnKYm-uWLh"
      },
      "source": [
        "### Interpreting Recommendation\n",
        "\n",
        "Below is the variance of each user latent dimension."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "01086B3oza3N",
        "outputId": "cd02454a-fd9e-42d7-e17c-e34081ec6bce",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 465
        }
      },
      "source": [
        "var_df = pd.DataFrame({\"Factor\": np.arange(K), \"Variance\": np.var(sorec.U, axis=0)})\n",
        "fig, ax = plt.subplots(figsize=(13, 5))\n",
        "sns.barplot(x=\"Factor\", y=\"Variance\", data=var_df, palette=\"ch:.25\", ax=ax);"
      ],
      "execution_count": 4,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1300x500 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "aizVEPmHuh8n"
      },
      "source": [
        "Let's select two dimensions and see how users are distributed in that 2D space.  In addition, we can visualize connections between two users if they are linked in the social network."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ndiIw1WJ8Ghv",
        "outputId": "54a764c1-4d59-4cd2-d2d2-4802cdf86662",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        }
      },
      "source": [
        "TOP2F = (9, 19)\n",
        "SAMPLE_SIZE = 200\n",
        "\n",
        "rng = np.random.RandomState(SEED)\n",
        "sample_inds = rng.choice(np.arange(sorec.U.shape[0]), size=SAMPLE_SIZE, replace=False)\n",
        "sample_df = pd.DataFrame(data=sorec.U[sample_inds][:, TOP2F], columns=[\"x\", \"y\"])\n",
        "g = sns.lmplot(x=\"x\", y=\"y\", data=sample_df, height=11.0, fit_reg=False)\n",
        "g.ax.set_title(\"Users in latent space with their social connections\", fontsize=16)\n",
        "\n",
        "adj_mat = sorec.train_set.user_graph.matrix\n",
        "for i in range(len(sample_inds)):\n",
        "  for j in range(len(sample_inds)):\n",
        "    if j != i and adj_mat[sample_inds[i], sample_inds[j]]:\n",
        "      sns.lineplot(x=\"x\", y=\"y\", data=sample_df.loc[[i, j]])"
      ],
      "execution_count": 5,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1100x1100 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "lNWQ6mDFu3uv"
      },
      "source": [
        "There are indeed some locality of user connections shown in the figure above. That suggests social network information is captured in the user latent factors and has some influence on the recommendations by the SoRec model. "
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "-8iXcfuKsVa8"
      },
      "source": [
        "## 3. Text Modality\n",
        "\n",
        "Often times, we are interested in building a recommender system for textual items (e.g., news, scientific papers), or items associated with text (e.g., titles, descriptions, reviews).  Text is informative and descriptive, therefore, exploiting textual information for better recommendations is an important topic in recommender systems.  In this tutorial, we introduce CTR [3], a recommendation model that combines matrix factorization and probablistic topic modeling. \n",
        "\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "742yJ53Fj1Zz"
      },
      "source": [
        "### Collaborative Topic Regression (CTR)\n",
        "\n",
        "Under factorization framework, adoption prediction is in the form of $\\hat{r}_{i,j} = \\mathbf{u}_i^T \\mathbf{v}_j $.  The intuition in CTR model is that two items with similar topics would behave similarly. Thus, item latent factors $\\mathbf{v_j}$ is assumed to be drawn from a Normal distribution:\n",
        "\n",
        "$$\n",
        "\\mathbf{v}_j \\sim \\mathcal{N}(\\mathbf{\\theta}_j, \\lambda^{-1} \\mathbf{I})\n",
        "$$\n",
        "\n",
        "where the mean $\\mathbf{\\theta}_j$ is a vector indicating topic proportions of the item $j$. It is equivalent to:\n",
        "\n",
        "\\begin{align}\n",
        "\\mathbf{v}_j &= \\mathbf{\\theta}_j + \\mathbf{\\epsilon}_j \\\\\n",
        "\\mathbf{\\epsilon}_j &\\sim \\mathcal{N}(\\mathbf{0}, \\lambda^{-1} \\mathbf{I})\n",
        "\\end{align}\n",
        "\n",
        "Please refer to paper [3] for the generative process of CTR model.\n",
        "\n",
        "\n",
        "CTR also extends matrix factorization, in which the base model is WMF under implicit feedback setting. The adoption $p_{i,j}$ and confidence $c_{i,j}$ are defined as follows: \n",
        "\n",
        "\\begin{equation}\n",
        "p_{i,j} = \n",
        "\\begin{cases} \n",
        "r_{i, j} &\\mbox{if } r_{i,j} > 0 \\\\\n",
        "0 & \\mbox{otherwise} \n",
        "\\end{cases}\n",
        "\\end{equation}\n",
        "\n",
        "\n",
        "\\begin{equation}\n",
        "c_{i,j} = \n",
        "\\begin{cases} \n",
        "a & \\mbox{if } r_{i,j} > 0 \\\\\n",
        "b & \\mbox{otherwise }\n",
        "\\end{cases}\n",
        "\\end{equation}\n",
        "\n",
        "The learning of CTR model is done via minimizing the following negative log-likelihood:\n",
        "\n",
        "$$ \\mathcal{L}(\\mathbf{U,V,\\theta, \\beta}|\\lambda) = \\frac{1}{2} \\sum_{i,j} c_{i,j} (p_{i,j} - \\mathbf{u}_i^T \\mathbf{v}_j)^2 - \\sum_{j}\\sum_{n} \\log \\big( \\sum_{k=1}^K \\mathbf{\\theta}_{j,k} \\mathbf{\\beta}_{k,w_{jn}} \\big) + \\frac{\\lambda}{2} \\sum_{i=1}^{N} ||\\mathbf{u}_i||^2 + \\frac{\\lambda}{2} \\sum_{j=1}^{M} (\\mathbf{v}_j - \\mathbf{\\theta}_j)^T (\\mathbf{v}_j - \\mathbf{\\theta}_j) $$\n",
        "\n",
        "It is an iterative procedure of alternating between three steps:\n",
        "- Optimize for user and item latent vectors, $\\mathbf{u}_i$ and $\\mathbf{v}_j$, based on the current topic proportions $\\mathbf{\\theta}_j$.  \n",
        "- Optimize for topic proportions $\\mathbf{\\theta}_j$ based on the current vectors $\\mathbf{u}_i$ and $\\mathbf{v}_j$ and topic words $\\mathbf{\\beta}_k$.\n",
        "- Optimize for topic words $\\mathbf{\\beta}_k$ based on the current topic proportions $\\mathbf{\\theta}_i$.\n",
        "\n",
        "Let's experiment with two models CTR and WMF on a dataset from Amazon Clothing category.  Using this dataset, CTR will learn topics from item description.  "
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "NZ2Al6omrlEP",
        "outputId": "972c38b4-7c97-4e4b-d93d-8a3f2bfdc494",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 367,
          "referenced_widgets": [
            "1f50664225f54fbdb5e900ef3279f97b",
            "f310aba5515d4b1eb74f3a0e422f2644",
            "1f13fb9c0979425aa03cfc9148f42dac",
            "a4052a5171e441dbb168fbacf865fb32",
            "f37087d5bf674d02ae75bcb5b75136d3",
            "3c835cfb9c8346df9a7f29784fd0c390",
            "69a532260693427b8985818cc9019203",
            "93819f2f8a1847f4911973d70b24248e",
            "21540e10e9ae4054bcfdf5a0c5820070",
            "f73928de4e764904a75e25f8fd4a4ac1",
            "199cf847db5d4d219a383abc28a8c6b3",
            "e10a12f760104a1ab3e181b33ec878bb",
            "342c1ec0286e496e87874f41102521d7",
            "faa730b609f84b60a58f07ac35290ebe",
            "39de41ba3b1645c3b4c748effa0f03af",
            "f2c786013d8f48b581d67f4e2edd898f",
            "1e8f16db87ff4a27945cc6fa9d053296",
            "1458df38c7f64297a88916fa52d5b960",
            "6cf235cfcea7430aadeaaf7e53c16de3",
            "847ca2b8872c40cc886bded5e076388b",
            "910379254e764589ba89499a00902aed",
            "a95b99387abf4d77bf51ce07f80b9d02"
          ]
        }
      },
      "source": [
        "K = 20\n",
        "ctr = CTR(k=K, max_iter=50, a=1.0, b=0.01, lambda_u=0.01, lambda_v=0.01, verbose=VERBOSE, seed=SEED)\n",
        "wmf = WMF(k=K, max_iter=50, a=1.0, b=0.01, learning_rate=0.005, lambda_u=0.01, lambda_v=0.01, \n",
        "          verbose=VERBOSE, seed=SEED)\n",
        "\n",
        "ratings = amazon_clothing.load_feedback()\n",
        "docs, item_ids = amazon_clothing.load_text()\n",
        "\n",
        "item_text_modality = TextModality(\n",
        "    corpus=docs,\n",
        "    ids=item_ids,\n",
        "    tokenizer=BaseTokenizer(sep=\" \", stop_words=\"english\"),\n",
        "    max_vocab=8000,\n",
        "    max_doc_freq=0.5,\n",
        ")\n",
        "\n",
        "ratio_split = RatioSplit(\n",
        "    data=ratings,\n",
        "    test_size=0.2,\n",
        "    rating_threshold=4.0,\n",
        "    exclude_unknowns=True,\n",
        "    item_text=item_text_modality,\n",
        "    verbose=VERBOSE,\n",
        "    seed=SEED,\n",
        ")\n",
        "\n",
        "rec_50 = cornac.metrics.Recall(50)\n",
        "\n",
        "cornac.Experiment(eval_method=ratio_split, models=[ctr, wmf], metrics=[rec_50]).run()"
      ],
      "execution_count": 6,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Data from https://static.preferred.ai/cornac/datasets/amazon_clothing/rating.zip\n",
            "will be cached into /root/.cornac/amazon_clothing/rating.txt\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "0.00B [00:00, ?B/s]"
            ],
            "application/vnd.jupyter.widget-view+json": {
              "version_major": 2,
              "version_minor": 0,
              "model_id": "1f50664225f54fbdb5e900ef3279f97b"
            }
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Unzipping ...\n",
            "File cached!\n",
            "Data from https://static.preferred.ai/cornac/datasets/amazon_clothing/text.zip\n",
            "will be cached into /root/.cornac/amazon_clothing/text.txt\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "0.00B [00:00, ?B/s]"
            ],
            "application/vnd.jupyter.widget-view+json": {
              "version_major": 2,
              "version_minor": 0,
              "model_id": "e10a12f760104a1ab3e181b33ec878bb"
            }
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Unzipping ...\n",
            "File cached!\n",
            "\n",
            "TEST:\n",
            "...\n",
            "    | Recall@50 | Train (s) | Test (s)\n",
            "--- + --------- + --------- + --------\n",
            "CTR |    0.2239 |  122.8733 |   1.0334\n",
            "WMF |    0.1582 |   15.1118 |   1.0026\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "uofgWV3oFLRq"
      },
      "source": [
        "The results show that CTR model performs significantly better than WMF model in terms of Recall@50, which is due to the contribution of items' textual information."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "r-8mCmABFe8e"
      },
      "source": [
        "### Interpreting Recommendation\n",
        "\n",
        "As CTR is based on topic model, it has intuitive explainability both in terms of model parameters as well as its recommendations. First, we can see what the top words of each topic are, based on learned topic-word distributions $\\mathbf{\\beta}$:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "fOY0pdwcC-VD",
        "outputId": "5404ec36-9b3f-49d9-9ef7-63fccae5162b",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 694
        }
      },
      "source": [
        "vocab = ctr.train_set.item_text.vocab\n",
        "topic_word_dist = ctr.model.beta.T[:, -ctr.train_set.item_text.max_vocab:] \n",
        "top_word_inds = np.argsort(topic_word_dist, axis=1) + 4  # ingore 4 special tokens\n",
        "\n",
        "topic_words = {}\n",
        "topic_df = defaultdict(list)\n",
        "print(\"WORD TOPICS:\")\n",
        "for t in range(len(topic_word_dist)):\n",
        "  top_words = vocab.to_text(top_word_inds[t][-10:][::-1], sep=\", \")\n",
        "  topic_words[t+1] = top_words\n",
        "  topic_df[\"Topic\"].append(t + 1)\n",
        "  topic_df[\"Top words\"].append(top_words)\n",
        "topic_df = pd.DataFrame(topic_df)\n",
        "topic_df"
      ],
      "execution_count": 7,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "WORD TOPICS:\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "    Topic  \\\n",
              "0   1       \n",
              "1   2       \n",
              "2   3       \n",
              "3   4       \n",
              "4   5       \n",
              "5   6       \n",
              "6   7       \n",
              "7   8       \n",
              "8   9       \n",
              "9   10      \n",
              "10  11      \n",
              "11  12      \n",
              "12  13      \n",
              "13  14      \n",
              "14  15      \n",
              "15  16      \n",
              "16  17      \n",
              "17  18      \n",
              "18  19      \n",
              "19  20      \n",
              "\n",
              "                                                                                  Top words  \n",
              "0   belt, flatlock, case, size, better, choose, earrings, soft, warm, collar                 \n",
              "1   shirt, style, breathable, princess, youre, colors, quality, come, people, socks          \n",
              "2   easy, collection, colors, rubber, water, quickly, stretch, lightweight, design, body     \n",
              "3   activities, handle, helmet, star, way, flat, makes, day, half, tee                       \n",
              "4   fit, sweat, mesh, large, pocket, fleece, body, necklace, imported, travel                \n",
              "5   wicks, material, dry, fit, panels, piece, insect, construction, durability, standard     \n",
              "6   support, company, fabric, apparel, fun, sizes, hip, outdoor, fiber, features             \n",
              "7   x, merino, sleeve, fun, items, water, weather, bracelets, comfortable, display           \n",
              "8   great, performance, black, pockets, inches, make, away, design, jewelry, unique          \n",
              "9   fits, sun, store, engineered, pocket, shield, conditions, mobility, hat, maximum         \n",
              "10  costume, products, small, inside, classic, lightweight, comfort, white, ideal, designed  \n",
              "11  bag, elastic, jacket, perfect, favorite, hat, d, box, look, feature                      \n",
              "12  s, use, sure, soft, comfort, easy, products, skin, displays, moisture                    \n",
              "13  size, comfortable, product, warm, need, ua, accessory, clean, footwear, silver           \n",
              "14  high, storage, height, waistband, wide, way, chest, include, growth, new                 \n",
              "15  delivers, tray, nylon, sturdy, free, pirate, builtin, allows, fairy, wear                \n",
              "16  cool, shoulder, provide, x, cold, prevents, womens, color, beautiful, years              \n",
              "17  pockets, mask, durable, moisture, jewelry, features, new, offers, comfort, brushed       \n",
              "18  straps, clothing, metal, way, microbes, shoes, silver, easily, toe, polyester            \n",
              "19  complete, wear, fabric, features, display, measures, merino, technology, gift, want      "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-eb289df5-21ba-4252-ba86-3a2eac7fa9f4\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>Topic</th>\n",
              "      <th>Top words</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>1</td>\n",
              "      <td>belt, flatlock, case, size, better, choose, earrings, soft, warm, collar</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>2</td>\n",
              "      <td>shirt, style, breathable, princess, youre, colors, quality, come, people, socks</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>3</td>\n",
              "      <td>easy, collection, colors, rubber, water, quickly, stretch, lightweight, design, body</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>4</td>\n",
              "      <td>activities, handle, helmet, star, way, flat, makes, day, half, tee</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>5</td>\n",
              "      <td>fit, sweat, mesh, large, pocket, fleece, body, necklace, imported, travel</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>5</th>\n",
              "      <td>6</td>\n",
              "      <td>wicks, material, dry, fit, panels, piece, insect, construction, durability, standard</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>6</th>\n",
              "      <td>7</td>\n",
              "      <td>support, company, fabric, apparel, fun, sizes, hip, outdoor, fiber, features</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>7</th>\n",
              "      <td>8</td>\n",
              "      <td>x, merino, sleeve, fun, items, water, weather, bracelets, comfortable, display</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>8</th>\n",
              "      <td>9</td>\n",
              "      <td>great, performance, black, pockets, inches, make, away, design, jewelry, unique</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>9</th>\n",
              "      <td>10</td>\n",
              "      <td>fits, sun, store, engineered, pocket, shield, conditions, mobility, hat, maximum</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>10</th>\n",
              "      <td>11</td>\n",
              "      <td>costume, products, small, inside, classic, lightweight, comfort, white, ideal, designed</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>11</th>\n",
              "      <td>12</td>\n",
              "      <td>bag, elastic, jacket, perfect, favorite, hat, d, box, look, feature</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>12</th>\n",
              "      <td>13</td>\n",
              "      <td>s, use, sure, soft, comfort, easy, products, skin, displays, moisture</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>13</th>\n",
              "      <td>14</td>\n",
              "      <td>size, comfortable, product, warm, need, ua, accessory, clean, footwear, silver</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>14</th>\n",
              "      <td>15</td>\n",
              "      <td>high, storage, height, waistband, wide, way, chest, include, growth, new</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>15</th>\n",
              "      <td>16</td>\n",
              "      <td>delivers, tray, nylon, sturdy, free, pirate, builtin, allows, fairy, wear</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>16</th>\n",
              "      <td>17</td>\n",
              "      <td>cool, shoulder, provide, x, cold, prevents, womens, color, beautiful, years</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>17</th>\n",
              "      <td>18</td>\n",
              "      <td>pockets, mask, durable, moisture, jewelry, features, new, offers, comfort, brushed</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>18</th>\n",
              "      <td>19</td>\n",
              "      <td>straps, clothing, metal, way, microbes, shoes, silver, easily, toe, polyester</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>19</th>\n",
              "      <td>20</td>\n",
              "      <td>complete, wear, fabric, features, display, measures, merino, technology, gift, want</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-eb289df5-21ba-4252-ba86-3a2eac7fa9f4')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-eb289df5-21ba-4252-ba86-3a2eac7fa9f4 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-eb289df5-21ba-4252-ba86-3a2eac7fa9f4');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 7
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "eHmSmoWnG8jX"
      },
      "source": [
        "As item vectors are close to its topic proportions, user vectors can also be used to explain what topics they are interested in."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "lTs0zjSmcEOK",
        "outputId": "4af5f87b-d884-4940-9744-d584cfd830cb",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 161
        }
      },
      "source": [
        "UIDX = 123\n",
        "TOPK = 5\n",
        "\n",
        "item_id2idx = ctr.train_set.iid_map\n",
        "item_idx2id = list(ctr.train_set.item_ids)\n",
        "\n",
        "print(f\"USER {UIDX} TOP-3 TOPICS:\")\n",
        "topic_df.loc[np.argsort(ctr.U[UIDX])[-3:][::-1]]"
      ],
      "execution_count": 8,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "USER 123 TOP-3 TOPICS:\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "    Topic  \\\n",
              "14  15      \n",
              "13  14      \n",
              "12  13      \n",
              "\n",
              "                                                                         Top words  \n",
              "14  high, storage, height, waistband, wide, way, chest, include, growth, new        \n",
              "13  size, comfortable, product, warm, need, ua, accessory, clean, footwear, silver  \n",
              "12  s, use, sure, soft, comfort, easy, products, skin, displays, moisture           "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-76d8a5ec-95a0-4685-9d46-26fcda0427c5\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>Topic</th>\n",
              "      <th>Top words</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>14</th>\n",
              "      <td>15</td>\n",
              "      <td>high, storage, height, waistband, wide, way, chest, include, growth, new</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>13</th>\n",
              "      <td>14</td>\n",
              "      <td>size, comfortable, product, warm, need, ua, accessory, clean, footwear, silver</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>12</th>\n",
              "      <td>13</td>\n",
              "      <td>s, use, sure, soft, comfort, easy, products, skin, displays, moisture</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-76d8a5ec-95a0-4685-9d46-26fcda0427c5')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-76d8a5ec-95a0-4685-9d46-26fcda0427c5 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-76d8a5ec-95a0-4685-9d46-26fcda0427c5');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 8
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "zkWRNVgyIR0r"
      },
      "source": [
        "Those are the topics most relevant to her interests. What will be recommendations to this user? "
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "7jHH_y_iH6Bt",
        "outputId": "456cc5f9-9364-4848-ea66-03118f52cb15",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 407
        }
      },
      "source": [
        "recommendations, scores = ctr.rank(UIDX)\n",
        "print(f\"\\nTOP {TOPK} RECOMMENDATIONS FOR USER {UIDX}:\")\n",
        "rec_df = defaultdict(list)\n",
        "for i in recommendations[:TOPK]:\n",
        "  rec_df[\"URL\"].append(f\"https://www.amazon.com/dp/{item_idx2id[i]}\")\n",
        "  rec_df[\"Description\"].append(ctr.train_set.item_text.corpus[i])\n",
        "pd.DataFrame(rec_df)"
      ],
      "execution_count": 9,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            "TOP 5 RECOMMENDATIONS FOR USER 123:\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "                                    URL  \\\n",
              "0  https://www.amazon.com/dp/B002M62ES6   \n",
              "1  https://www.amazon.com/dp/B002MKPKFG   \n",
              "2  https://www.amazon.com/dp/B001NAHW2U   \n",
              "3  https://www.amazon.com/dp/B000HS6NEC   \n",
              "4  https://www.amazon.com/dp/B001EDF3OU   \n",
              "\n",
              "                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               Description  \n",
              "0  Accessorize your Spider-man Costume with these Gloves. Fits most child sizes from 4-12.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                  \n",
              "1  This glamorous dress is perfect for a night at the ball. The stretchy blue velvet bodice is elegantly detailed with a gathered white inlay, ruffled cap sleeves, sparkling silver trim and a white rose. The silver trimmed waist drapes outlined with ruffle capture that authentic Cinderella look.The dress is designed with an empire waist that sits 1.5 inches above the waist. They are machine washable and made for long repeated use, unlike many one-off Halloween costumes that are not made for everyday play.  Our stretchy accommodating fabrics will allow for a comfortable fit for each child within the respective age ranges and there are no itchy fabrics or trims.Each dress comes with an exclusive Princess Hairbow as part of the bundle set.  \n",
              "2  LACY 14K GOLD on 925 Silver Charm BEAD compatible with European Bracelet                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 \n",
              "3  Includes jacket, elastic pants, hat, belt, whistle, and gun holster. Child Size Small (4-6).                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                             \n",
              "4  The Child Indiana Jones Hat and Whip are two must have accessories for your child's Officially Licensed Indiana Jones costume. This Indiana Jones costume accessory contains the two hallmarks of Indy's style: his awesome whip and his trademark hat.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-a72cffe3-6ea3-4130-83ea-fabdbb657543\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>URL</th>\n",
              "      <th>Description</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>https://www.amazon.com/dp/B002M62ES6</td>\n",
              "      <td>Accessorize your Spider-man Costume with these Gloves. Fits most child sizes from 4-12.</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>https://www.amazon.com/dp/B002MKPKFG</td>\n",
              "      <td>This glamorous dress is perfect for a night at the ball. The stretchy blue velvet bodice is elegantly detailed with a gathered white inlay, ruffled cap sleeves, sparkling silver trim and a white rose. The silver trimmed waist drapes outlined with ruffle capture that authentic Cinderella look.The dress is designed with an empire waist that sits 1.5 inches above the waist. They are machine washable and made for long repeated use, unlike many one-off Halloween costumes that are not made for everyday play.  Our stretchy accommodating fabrics will allow for a comfortable fit for each child within the respective age ranges and there are no itchy fabrics or trims.Each dress comes with an exclusive Princess Hairbow as part of the bundle set.</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>https://www.amazon.com/dp/B001NAHW2U</td>\n",
              "      <td>LACY 14K GOLD on 925 Silver Charm BEAD compatible with European Bracelet</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>https://www.amazon.com/dp/B000HS6NEC</td>\n",
              "      <td>Includes jacket, elastic pants, hat, belt, whistle, and gun holster. Child Size Small (4-6).</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>https://www.amazon.com/dp/B001EDF3OU</td>\n",
              "      <td>The Child Indiana Jones Hat and Whip are two must have accessories for your child's Officially Licensed Indiana Jones costume. This Indiana Jones costume accessory contains the two hallmarks of Indy's style: his awesome whip and his trademark hat.</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-a72cffe3-6ea3-4130-83ea-fabdbb657543')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-a72cffe3-6ea3-4130-83ea-fabdbb657543 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-a72cffe3-6ea3-4130-83ea-fabdbb657543');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 9
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "L2E8LxyoIuZa"
      },
      "source": [
        "Interestingly, in the desctiption of top-5 recommendations, we see words relevant to the user's topics (e.g., *Accessorize, Gloves, Bracelet, Silver*). This would suggest that CTR model can capture user preferences via the notions of item topics. "
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "0jb-o9QJKcaS"
      },
      "source": [
        "---\n",
        "\n",
        "Let's check out another user with *index = 321*. Similarly, we see the top-3 topics this user cares most about: "
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "aKowd8UQQjkT",
        "outputId": "ba7d23d7-37a3-499b-ba8b-2e7a07cee852",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 161
        }
      },
      "source": [
        "UIDX = 321\n",
        "\n",
        "print(f\"USER {UIDX} TOP-3 TOPICS:\")\n",
        "topic_df.loc[np.argsort(ctr.U[UIDX])[-3:][::-1]]"
      ],
      "execution_count": 10,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "USER 321 TOP-3 TOPICS:\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "    Topic  \\\n",
              "19  20      \n",
              "2   3       \n",
              "10  11      \n",
              "\n",
              "                                                                                  Top words  \n",
              "19  complete, wear, fabric, features, display, measures, merino, technology, gift, want      \n",
              "2   easy, collection, colors, rubber, water, quickly, stretch, lightweight, design, body     \n",
              "10  costume, products, small, inside, classic, lightweight, comfort, white, ideal, designed  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-a05098b8-96ff-4f55-8d08-5278c324705a\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>Topic</th>\n",
              "      <th>Top words</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>19</th>\n",
              "      <td>20</td>\n",
              "      <td>complete, wear, fabric, features, display, measures, merino, technology, gift, want</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>3</td>\n",
              "      <td>easy, collection, colors, rubber, water, quickly, stretch, lightweight, design, body</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>10</th>\n",
              "      <td>11</td>\n",
              "      <td>costume, products, small, inside, classic, lightweight, comfort, white, ideal, designed</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-a05098b8-96ff-4f55-8d08-5278c324705a')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-a05098b8-96ff-4f55-8d08-5278c324705a button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-a05098b8-96ff-4f55-8d08-5278c324705a');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 10
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "29isliv_K0uf"
      },
      "source": [
        "The topics include clothing with certain characteristics (*easy, water, stretch, lightweight, comfort*).\n",
        "\n",
        "Based on this, how would the model give recommendations to the user?"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "z7mvxNntLo8V",
        "outputId": "ce1488a3-7f46-44f8-9e1c-398d49a5891b",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 564
        }
      },
      "source": [
        "recommendations, scores = ctr.rank(UIDX)\n",
        "print(f\"\\nTOP {TOPK} RECOMMENDATIONS FOR USER {UIDX}:\")\n",
        "rec_df = defaultdict(list)\n",
        "for i in recommendations[:TOPK]:\n",
        "  rec_df[\"URL\"].append(f\"https://www.amazon.com/dp/{item_idx2id[i]}\")\n",
        "  rec_df[\"Description\"].append(ctr.train_set.item_text.corpus[i])\n",
        "pd.DataFrame(rec_df)"
      ],
      "execution_count": 11,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            "TOP 5 RECOMMENDATIONS FOR USER 321:\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "                                    URL  \\\n",
              "0  https://www.amazon.com/dp/B000MARQFA   \n",
              "1  https://www.amazon.com/dp/B001VIH648   \n",
              "2  https://www.amazon.com/dp/B002V44LTO   \n",
              "3  https://www.amazon.com/dp/B002FIVZD6   \n",
              "4  https://www.amazon.com/dp/B00062B1IC   \n",
              "\n",
              "                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                   Description  \n",
              "0  You demand a lot from your pants when you're on the move and the PROPPER Tactical Pant delivers. Nine pockets hold everything from your smartphone to your multi-tool. Rugged canvas fabric offers durability and strength, with DuPont Teflon fabric protector to repel stains and liquids. An action-stretch waistband adds comfort. And with our innovative wallet \"pocket in a pocket\" you'll appreciate not sitting on your wallet all day long. Come with a free matching belt (12.99 value).                                                                                                                          \n",
              "1  You demand a lot from your pants when you're on the move and the PROPPER Tactical Pant delivers. Nine pockets hold everything from your smartphone to your multi-tool. Rugged canvas fabric offers durability and strength, with DuPont Teflon fabric protector to repel stains and liquids. An action-stretch waistband adds comfort. And with our innovative wallet \"pocket in a pocket\" you'll appreciate not sitting on your wallet all day long. Come with a free matching belt (12.99 value).                                                                                                                          \n",
              "2  Sleeves - Made From 50% nylon & 50% cotton. Chest & Back - Made from 40% Acrylic & 60% Cotton. Resists melting & dripping. Melting point is rated to 230 degrees Fahrenheit. Moisture wicking. Lightweight & breathable. Zippered sleeve storage pockets. Mock turtle neck. Anti abrasion padded elbow patches. Gusseted sleeves. Glint tape holder for IR tape. Loop on sleeve pockets for attaching name, rank and flag. Colors Available: Black, ACU Digital Camouflage, Olive Drab, Desert Digital Camouflage & Digitial Woodland Camouflage. Sizes to choose from: Small, Medium, Large, X-Large, 2X-Large & 3X-Large.  \n",
              "3  TRU-SPEC is one of the leading suppliers of uniforms and personal equipment to military, law enforcement and public safety markets.&#xA0; Thousands of police, firefighters, EMTs and military personnel use and depend on our products daily. The name TRU-SPEC has become synonymous with quality, innovation and service. Gunnery Sgt. R. Lee Ermey, USMC (Ret.) known as \"The Gunny\" has become the official celebrity spokesperson for TRU-SPEC's line of tactically inspired apparel.                                                                                                                                  \n",
              "4  The PROPPER BDU trouser offers an authentic military cut in comfortable, durable Battle Rip fabric. Sewn to military specification, this button-fly BDU is fade, shrink and wrinkle resistant. It features felled seams and reinforced seat and knee for maximum durability. Six pockets with fused pocket flaps ensure a clean, professional look. With adjustable waist tabs and durable drawstring leg closures for a secure fit.                                                                                                                                                                                         "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-751e2e57-d491-4ddb-bb80-1055eb74c63b\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>URL</th>\n",
              "      <th>Description</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>https://www.amazon.com/dp/B000MARQFA</td>\n",
              "      <td>You demand a lot from your pants when you're on the move and the PROPPER Tactical Pant delivers. Nine pockets hold everything from your smartphone to your multi-tool. Rugged canvas fabric offers durability and strength, with DuPont Teflon fabric protector to repel stains and liquids. An action-stretch waistband adds comfort. And with our innovative wallet \"pocket in a pocket\" you'll appreciate not sitting on your wallet all day long. Come with a free matching belt (12.99 value).</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>https://www.amazon.com/dp/B001VIH648</td>\n",
              "      <td>You demand a lot from your pants when you're on the move and the PROPPER Tactical Pant delivers. Nine pockets hold everything from your smartphone to your multi-tool. Rugged canvas fabric offers durability and strength, with DuPont Teflon fabric protector to repel stains and liquids. An action-stretch waistband adds comfort. And with our innovative wallet \"pocket in a pocket\" you'll appreciate not sitting on your wallet all day long. Come with a free matching belt (12.99 value).</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>https://www.amazon.com/dp/B002V44LTO</td>\n",
              "      <td>Sleeves - Made From 50% nylon &amp; 50% cotton. Chest &amp; Back - Made from 40% Acrylic &amp; 60% Cotton. Resists melting &amp; dripping. Melting point is rated to 230 degrees Fahrenheit. Moisture wicking. Lightweight &amp; breathable. Zippered sleeve storage pockets. Mock turtle neck. Anti abrasion padded elbow patches. Gusseted sleeves. Glint tape holder for IR tape. Loop on sleeve pockets for attaching name, rank and flag. Colors Available: Black, ACU Digital Camouflage, Olive Drab, Desert Digital Camouflage &amp; Digitial Woodland Camouflage. Sizes to choose from: Small, Medium, Large, X-Large, 2X-Large &amp; 3X-Large.</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>https://www.amazon.com/dp/B002FIVZD6</td>\n",
              "      <td>TRU-SPEC is one of the leading suppliers of uniforms and personal equipment to military, law enforcement and public safety markets.&amp;#xA0; Thousands of police, firefighters, EMTs and military personnel use and depend on our products daily. The name TRU-SPEC has become synonymous with quality, innovation and service. Gunnery Sgt. R. Lee Ermey, USMC (Ret.) known as \"The Gunny\" has become the official celebrity spokesperson for TRU-SPEC's line of tactically inspired apparel.</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>https://www.amazon.com/dp/B00062B1IC</td>\n",
              "      <td>The PROPPER BDU trouser offers an authentic military cut in comfortable, durable Battle Rip fabric. Sewn to military specification, this button-fly BDU is fade, shrink and wrinkle resistant. It features felled seams and reinforced seat and knee for maximum durability. Six pockets with fused pocket flaps ensure a clean, professional look. With adjustable waist tabs and durable drawstring leg closures for a secure fit.</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-751e2e57-d491-4ddb-bb80-1055eb74c63b')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-751e2e57-d491-4ddb-bb80-1055eb74c63b button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-751e2e57-d491-4ddb-bb80-1055eb74c63b');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 11
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "1xQu4pblNZRq"
      },
      "source": [
        "From the top-5 recommendations, we notice that those items are *military/outdoor/sports* clothing, based on their descriptions as well as the URL to Amazon website. That is also evident from the user's topics that we observe above."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "q1AQfgGe8fyo"
      },
      "source": [
        "## 4. Image Modality\n",
        "\n",
        "In some contexts, item images are informative (e.g., fashion). With the existence of effective methods to learn image representation, using item images in recommender systems is gaining popularity. In this tutorial, we present VBPR [4], a recommendation model making use of item image features extracted from pre-trained Convolutional Neural Network (CNN)."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "A7WGq2xxjqjA"
      },
      "source": [
        "### Visual Bayesian Personalized Ranking (VBPR)\n",
        "\n",
        "VBPR, which is also based on matrix factorization, is an extension of BPR model.  The novelty of VBPR is on how item visual features incorporated into the matrix factorization framework.  The preference score user $i$ giving to item $j$ is predicted as follows:\n",
        "\n",
        "$$\n",
        "\\hat{r}_{i,j} = \\alpha + b_i + b_j + \\mathbf{u}_i^T \\mathbf{v}_j + \\mathbf{p}_{i}^T(\\mathbf{E} \\times \\mathbf{f}_j) + \\mathbf{\\Theta}^T \\mathbf{f}_j\n",
        "$$\n",
        "\n",
        "where:\n",
        "- $\\alpha, b_i, b_j$ are global bias, user bias, and item bias, respectively\n",
        "- $\\mathbf{u}_i \\in \\mathbb{R}^K$ and $\\mathbf{v}_j \\in \\mathbb{R}^K$ are user and item latent vectors, respectively\n",
        "- $\\mathbf{f}_j \\in \\mathbb{R}^D$ is the item image feature vector\n",
        "- $\\mathbf{p}_i \\in \\mathbb{R}^Q$ is user visual preference, and $(\\mathbf{E} \\times \\mathbf{f}_j) \\in \\mathbb{R}^Q$ is item visual representation with $\\mathbf{E} \\in \\mathbb{R}^{K \\times D}$ is the projection from visual feature space into preference space\n",
        "- $\\mathbf{\\Theta} \\in \\mathbb{R}^D$ is global visual bias vector\n",
        "\n",
        "Learning parameters of VBPR model can be done, similarly to BPR, via minimizing the following negative log-likelihood:\n",
        "\n",
        "$$ \\mathcal{L}(\\mathbf{U,V,b,E,\\Theta, P}|\\lambda) = \\sum_{(j >_i l) \\in \\mathbf{S}} \\ln (1 + \\exp\\{-(\\hat{r}_{i,j} - \\hat{r}_{i,l})\\}) + \\frac{\\lambda}{2} \\sum_{i=1}^{N} (||\\mathbf{u}_i||^2 + ||\\mathbf{p}_i||^2) + \\frac{\\lambda}{2} \\sum_{j=1}^{M} (b_j + ||\\mathbf{v}_j||^2) + \\frac{\\lambda}{2} ||\\mathbf{\\Theta}||^2 + \\frac{\\lambda}{2} ||\\mathbf{E}||^2_2 $$\n",
        "\n",
        "Noted that global bias $\\alpha$ and user bias $b_i$ do not affect the ranking of items, thus they are redundant and removed from the model parameters.\n",
        "\n",
        "Let's compare VBPR and BPR models with an experiment on Amazon Clothing dataset.\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "4LQDDr1e8j6L",
        "outputId": "5afa001b-f291-4bb9-8da4-88c90d6d2b8c",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 385,
          "referenced_widgets": [
            "04c0fda0b1b14d85973e24befbd0997b",
            "fe5af89a90b6460aa992b62938552d7b",
            "7efbd2a55a7244a89c007269e067ec9a",
            "c505ce61346d4dfb9c214322ac955b56",
            "6b7afb98961a4a918ddb83879409046d",
            "f7ad5692ade94c87be902a84acff243d",
            "8156dc94e21940be9a556b06e5f818f8",
            "18d7c30049f048f8a27bda6bf3a3b220",
            "7aeb2766deb44ff1a4b04930f5b2c7e1",
            "77ae7062d88f4e5db5cd2b4fd917825c",
            "3102c018ee354b2ba0be7389bca3519e",
            "640728869a334399b926a0c1e3236c81",
            "94c1106c817f41368c4e42d5455254d4",
            "9b5f169748fd40a1b92502afe2592ac3",
            "588760830bfb45d9b76b64ae9370988b",
            "f04e7f82eb03434a9a6755197cb906a7",
            "aac8895239cc4100854c16b36ca89f59",
            "292153a6e05d47d49cc901141aa7e347",
            "9e531d3678be4fa38f6467d562ac5286",
            "3d7e45fb56914571b447bd28f04ec470",
            "7c94da4bbef3481aac2d28bbd165154f",
            "41ed6a1c3b0a49ee9ab22a91f8659797"
          ]
        }
      },
      "source": [
        "K = 10\n",
        "vbpr = VBPR(k=K, k2=K, n_epochs=50, batch_size=100, learning_rate=0.001,\n",
        "            lambda_w=1, lambda_b=0.01, lambda_e=0.0, use_gpu=True, verbose=VERBOSE, seed=SEED)\n",
        "bpr = BPR(k=(K * 2), max_iter=50, learning_rate=0.001, lambda_reg=0.001, verbose=VERBOSE, seed=SEED)\n",
        "\n",
        "ratings = amazon_clothing.load_feedback()\n",
        "img_features, item_ids = amazon_clothing.load_visual_feature()\n",
        "\n",
        "item_image_modality = ImageModality(features=img_features, ids=item_ids, normalized=True)\n",
        "\n",
        "ratio_split = RatioSplit(\n",
        "    data=ratings,\n",
        "    test_size=0.2,\n",
        "    rating_threshold=4.0,\n",
        "    exclude_unknowns=True,\n",
        "    item_image=item_image_modality,\n",
        "    verbose=VERBOSE,\n",
        "    seed=SEED,\n",
        ")\n",
        "\n",
        "auc = cornac.metrics.AUC()\n",
        "\n",
        "cornac.Experiment(eval_method=ratio_split, models=[vbpr, bpr], metrics=[auc]).run()"
      ],
      "execution_count": 12,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Data from https://static.preferred.ai/cornac/datasets/amazon_clothing/image.zip\n",
            "will be cached into /root/.cornac/amazon_clothing/image_features.npy\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "0.00B [00:00, ?B/s]"
            ],
            "application/vnd.jupyter.widget-view+json": {
              "version_major": 2,
              "version_minor": 0,
              "model_id": "04c0fda0b1b14d85973e24befbd0997b"
            }
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Unzipping ...\n",
            "File cached!\n",
            "Data from https://static.preferred.ai/cornac/datasets/amazon_clothing/item_ids.zip\n",
            "will be cached into /root/.cornac/amazon_clothing/item_ids.txt\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "0.00B [00:00, ?B/s]"
            ],
            "application/vnd.jupyter.widget-view+json": {
              "version_major": 2,
              "version_minor": 0,
              "model_id": "640728869a334399b926a0c1e3236c81"
            }
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Unzipping ...\n",
            "File cached!\n",
            "Optimization finished!\n",
            "\n",
            "TEST:\n",
            "...\n",
            "     |    AUC | Train (s) | Test (s)\n",
            "---- + ------ + --------- + --------\n",
            "VBPR | 0.7053 |   39.8092 |   1.1703\n",
            "BPR  | 0.5992 |    0.0919 |   0.9641\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "MR6coclHv1Do"
      },
      "source": [
        "The results show that VBPR obtains higher performance than BPR in terms of AUC. That can be attributed to the usage of item visual features. Let's try to get some ideas on how the visual features contributing to the model recommendations. "
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "JAdxJSASpswm"
      },
      "source": [
        "### Interpreting Recommendation\n",
        "\n",
        "First, let's take a look at items rated by a particular user. Then we see how VBPR model recommends to the user."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "i3FxEMm75Y_R",
        "outputId": "716da4b2-3246-486a-e55f-653fd80c2c5e",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 130
        }
      },
      "source": [
        "UIDX = 111\n",
        "\n",
        "item_id2idx = vbpr.train_set.iid_map\n",
        "item_idx2id = list(vbpr.train_set.item_ids)\n",
        "img_features = vbpr.train_set.item_image.features\n",
        "item_desc = {iid: desc for desc, iid in zip(*amazon_clothing.load_text())}\n",
        "\n",
        "print(f\"RATED ITEMS BY USER {UIDX}:\")\n",
        "rated_items = np.nonzero(vbpr.train_set.matrix[UIDX].A)[1]\n",
        "rated_df = defaultdict(list)\n",
        "for i in rated_items:\n",
        "  rated_df[\"URL\"].append(f\"https://www.amazon.com/dp/{item_idx2id[i]}\")\n",
        "  rated_df[\"Description\"].append(item_desc[item_idx2id[i]])\n",
        "pd.DataFrame(rated_df)"
      ],
      "execution_count": 13,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "RATED ITEMS BY USER 111:\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "                                    URL  \\\n",
              "0  https://www.amazon.com/dp/B00BXQF1XG   \n",
              "1  https://www.amazon.com/dp/B002M62ES6   \n",
              "\n",
              "                                                                               Description  \n",
              "0  Disguise Boy's Marvel Ultimate Spider-Man Classic Costume, 3T-4T Jumpsuit.               \n",
              "1  Accessorize your Spider-man Costume with these Gloves. Fits most child sizes from 4-12.  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-2492bcb9-81b2-46f9-ad58-e87335e07f45\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>URL</th>\n",
              "      <th>Description</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>https://www.amazon.com/dp/B00BXQF1XG</td>\n",
              "      <td>Disguise Boy's Marvel Ultimate Spider-Man Classic Costume, 3T-4T Jumpsuit.</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>https://www.amazon.com/dp/B002M62ES6</td>\n",
              "      <td>Accessorize your Spider-man Costume with these Gloves. Fits most child sizes from 4-12.</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-2492bcb9-81b2-46f9-ad58-e87335e07f45')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-2492bcb9-81b2-46f9-ad58-e87335e07f45 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-2492bcb9-81b2-46f9-ad58-e87335e07f45');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 13
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "5BA12M75R6W_",
        "outputId": "7e64fe2e-d3b2-4050-d1cf-2a58d6fe2344",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 529
        }
      },
      "source": [
        "TOPK = 5\n",
        "\n",
        "recommendations, scores = vbpr.rank(UIDX)\n",
        "\n",
        "print(f\"\\nTOP {TOPK} RECOMMENDATIONS:\")\n",
        "rec_df = defaultdict(list)\n",
        "for i in recommendations[:TOPK]:\n",
        "  cf_contrib = vbpr.beta_item[i] + np.dot(vbpr.gamma_item[i], vbpr.gamma_user[UIDX])\n",
        "  vis_contrib = vbpr.visual_bias[i] + np.dot(vbpr.theta_item[i], vbpr.theta_user[UIDX])\n",
        "  rec_df[\"URL\"].append(f\"https://www.amazon.com/dp/{item_idx2id[i]}\")\n",
        "  rec_df[\"Score\"].append(scores[i])\n",
        "  rec_df[\"CF Contribution\"].append(cf_contrib)\n",
        "  rec_df[\"Visual Contribution\"].append(vis_contrib)\n",
        "  rec_df[\"Description\"].append(item_desc[item_idx2id[i]])\n",
        "pd.DataFrame(rec_df)"
      ],
      "execution_count": 14,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            "TOP 5 RECOMMENDATIONS:\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "                                    URL     Score  CF Contribution  \\\n",
              "0  https://www.amazon.com/dp/B002M62ES6  7.457715  1.116012          \n",
              "1  https://www.amazon.com/dp/B007DLVLAW  7.312137  1.509102          \n",
              "2  https://www.amazon.com/dp/B00BHPRRKS  6.589272 -0.176793          \n",
              "3  https://www.amazon.com/dp/B0000CNR1L  6.477899  0.324168          \n",
              "4  https://www.amazon.com/dp/B00BXQF1XG  6.426818  0.020040          \n",
              "\n",
              "   Visual Contribution  \\\n",
              "0  6.341703              \n",
              "1  5.803034              \n",
              "2  6.766065              \n",
              "3  6.153731              \n",
              "4  6.406778              \n",
              "\n",
              "                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                  Description  \n",
              "0  Accessorize your Spider-man Costume with these Gloves. Fits most child sizes from 4-12.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                     \n",
              "1  Patented elastic shoelace and fastening system. Engineered and designed to improve performance and keep laces locked and secure throughout the course of a race, game, practice, or workout.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                \n",
              "2  Jumpsuit with muscle torso and arms attached belt.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                          \n",
              "3  No more digging around in deep toiletries bags for that pair of tweezers or spare safety pin you know is in there somewhere. Made of clear vinyl for quick viewing, the ten compartments on this unique organizer keep everything from dental floss to shaving cream readily retrievable. Organized in four zippered rows, the compartments vary in size and keep contents secure. When empty, the bag folds to a flat 10 by 14 inches. When filled, it can be laid inside a suitcase. The hook allows the bag to be hung inside a garment bag if desired, and then looped over a bathroom door or towel hook for accessibility.  Made of PVC microfiber, the bag extends to 28 by 14 inches when fully opened.--Ann Bieri  \n",
              "4  Disguise Boy's Marvel Ultimate Spider-Man Classic Costume, 3T-4T Jumpsuit.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-d4d5cd03-0a0e-49a1-aed2-5dd71e2c9ff8\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>URL</th>\n",
              "      <th>Score</th>\n",
              "      <th>CF Contribution</th>\n",
              "      <th>Visual Contribution</th>\n",
              "      <th>Description</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>https://www.amazon.com/dp/B002M62ES6</td>\n",
              "      <td>7.457715</td>\n",
              "      <td>1.116012</td>\n",
              "      <td>6.341703</td>\n",
              "      <td>Accessorize your Spider-man Costume with these Gloves. Fits most child sizes from 4-12.</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>https://www.amazon.com/dp/B007DLVLAW</td>\n",
              "      <td>7.312137</td>\n",
              "      <td>1.509102</td>\n",
              "      <td>5.803034</td>\n",
              "      <td>Patented elastic shoelace and fastening system. Engineered and designed to improve performance and keep laces locked and secure throughout the course of a race, game, practice, or workout.</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>https://www.amazon.com/dp/B00BHPRRKS</td>\n",
              "      <td>6.589272</td>\n",
              "      <td>-0.176793</td>\n",
              "      <td>6.766065</td>\n",
              "      <td>Jumpsuit with muscle torso and arms attached belt.</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>https://www.amazon.com/dp/B0000CNR1L</td>\n",
              "      <td>6.477899</td>\n",
              "      <td>0.324168</td>\n",
              "      <td>6.153731</td>\n",
              "      <td>No more digging around in deep toiletries bags for that pair of tweezers or spare safety pin you know is in there somewhere. Made of clear vinyl for quick viewing, the ten compartments on this unique organizer keep everything from dental floss to shaving cream readily retrievable. Organized in four zippered rows, the compartments vary in size and keep contents secure. When empty, the bag folds to a flat 10 by 14 inches. When filled, it can be laid inside a suitcase. The hook allows the bag to be hung inside a garment bag if desired, and then looped over a bathroom door or towel hook for accessibility.  Made of PVC microfiber, the bag extends to 28 by 14 inches when fully opened.--Ann Bieri</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>https://www.amazon.com/dp/B00BXQF1XG</td>\n",
              "      <td>6.426818</td>\n",
              "      <td>0.020040</td>\n",
              "      <td>6.406778</td>\n",
              "      <td>Disguise Boy's Marvel Ultimate Spider-Man Classic Costume, 3T-4T Jumpsuit.</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-d4d5cd03-0a0e-49a1-aed2-5dd71e2c9ff8')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-d4d5cd03-0a0e-49a1-aed2-5dd71e2c9ff8 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-d4d5cd03-0a0e-49a1-aed2-5dd71e2c9ff8');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 14
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "G78hw2wTk5q-"
      },
      "source": [
        "The recommendations (0, 2, 4), with high visual contribution, are indeed visually similar to the two items rated by the user.  That gives us some intuition on how VPBR model makes use of the item visual features in the predictions."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "0FmiuTYiP_pB"
      },
      "source": [
        "## References\n",
        "\n",
        "1.   Aggarwal, C. C. (2016). Recommender systems (Vol. 1). Cham: Springer International Publishing.\n",
        "2.   Ma, H., Yang, H., Lyu, M. R., & King, I. (2008, October). Sorec: social recommendation using probabilistic matrix factorization. In Proceedings of the 17th ACM conference on Information and knowledge management (pp. 931-940).\n",
        "3.   Wang, C., & Blei, D. M. (2011, August). Collaborative topic modeling for recommending scientific articles. In Proceedings of the 17th ACM SIGKDD international conference on Knowledge discovery and data mining (pp. 448-456).   \n",
        "4.   He, R., & McAuley, J. (2016, February). VBPR: visual bayesian personalized ranking from implicit feedback. In Thirtieth AAAI Conference on Artificial Intelligence.\n",
        "5.   Cornac - A Comparative Framework for Multimodal Recommender Systems (https://cornac.preferred.ai/)\n",
        "\n"
      ]
    }
  ]
}