Explanations come in two forms: local, explaining a single model prediction, and global, explaining all model predictions. The Local to Global (L2G) problem consists of bridging these two families of explanations. Simply put, we generate global explanations by merging local ones.
FLocalX is an open source Python Library that provides a framework to explore the creation of global explanations derived from local explanations in the form of rulesets. The objective of the library is to be extensible with new explainers and metaheuristics approaches to create new global explanations.
FLocalX requires:
* Python (>=3.9)
* NumPy
* Scikit-Learn
* Scikit-fuzzy
IMPORTANT Install scikit-fuzzy from their GitHub as the PyPi version is obsolete:
pip install git+https://github.com/scikit-fuzzy/scikit-fuzzyIf you already have a working installation, you can install FLocalX with
git clone https://github.com/Kaysera/flocalx
pip install flocalxFor detailed instructions on how to use FLocalX, please refer to the examples folder
The following explainers are currently supported:
- LORE: Local explainer generated from a neighborhood
- FLARE: Fuzzy local explainer generated from a neighborhood
The following metaheuristics are currently supported:
- Genetic Algorithm
- FLocalX (accepted at IDA 2024)
- Fuzzy Factuals and counterfactuals(Fernandez et al., 2022)
- Documentation https://xai-teacher.readthedocs.io/en/latest/
- Experiments: https://github.com/Kaysera/teacher-experiments
- LORE (Guidotti et al., 2018)
- Documentation and examples: https://doi.org/10.1109/MIS.2019.2957223
- FLARE (Fernandez et al., 2023 preprint)