gale

gale provides functions to globally compare local explanations. To install the library, run pip install gale-topo.

GALE is a Python library used to assess the similarity of local explanations from methods such as LIME, SHAP or generalized additive models (GAMs). To do so, GALE models the relationship between the explanation space and the model predictions as a scalar function. Then, we compute the topological skeleton of this function. This topological skeleton acts as a signature, which we use to compare outputs from different explanation methods. GALE is easy to use and can be deployed in just a few lines of code, as we show below

from gale import create_mapper, bottleneck_distance

model_one_mapper = create_mapper(explanations_one, predictions_one, resolution=10, gain=0.3, dist_thresh=0.5)
model_two_mapper = create_mapper(explanations_two, predictions_two, resolution=10, gain=0.3, dist_thresh=0.5)

bottleneck_distance(model_one_mapper, model_two_mapper)

Using gale

Installation

How to install gale.

Examples

Examples to help get you started with GALE.

Frequently Asked Questions (FAQs)

Need help? Check the FAQs first.

License and Acknowledgments

License and acknowledgments.

gale Modules

Mapper

Mapper module.