Schedule.md

Deep Learning Bootcamp

Session 1: Deep Learning (DL) Theory

1. What is deep learning?
2. Use cases in computer vision and natural language processing.
3. Introduction to problem and data-set - Fashion MNIST
4. Working on the cloud, including keras and tensorflow: Rorodeep
5. Build your ML model - t-SNE + Multi-Class Logistic Regression (sklearn)
6. Build your first DL Model - Multi-layer Perceptron (MLP)

Session 2: Overview of the building blocks

1. Input, Output & Loss Function
2. Neurons
3. Activation functions
4. Back propagation algorithm
5. Stochastic gradient descent / Adaptive learning / Momentum
6. Training & Validation
7. Architecture

Session 3: Convolutional Neural Networks (CNN)

1. Build your second DL Model - CNN
2. Concept of Convolution
   • Kernel Size
   • Shared Weights
   • Pooling
   • Padding & Stride
3. Tricks to improve your model
   • Augment your training data
   • Batch normalization (Kill)

Session 4: Transfer Learning

1. Concept of Transfer Learning
2. Build your third DL Model - Leverage pre-trained models
3. Deploying your DL model on the cloud (Flask and deploy - Jinja template)

Session 5: DL for Natural Language Processing (NLP)

1. Recap of Day One
2. Challanges with traditional NLP techniques
3. Concept of Word Embedding - word2vec
4. Build your fourth DL Model - MLP using word2vec

Session 6: Recurrent Neural Networks (RNN)

1. Concept of RNNs
2. Concept of Long Short-Term Memory (LSTM)
3. Build your fifth DL Model - LSTM

Session 7: Build your DL Applications

1. Concept of Sequence-to-Sequence Learning
2. Build your sixth DL Model - Text Generation
3. Deploy it as a bot (e.g. TweetBot / ChatBot)

Session 8: Advanced Topics in DL (Theory)

1. Challenges in building DL apps
2. Concept of Generative Adversarial Network
3. Moving beyond Classification e.g. Object Detection
4. Concept of DL for Unsupervised Learning
5. Concept of Reinforcement Learning
6. Where to go from here...