Xai Library
Our Python library designed to develop explainable machine learning models.
The library provides an integrated pipeline to set up and execute explanation methods for black boxes.
ISIC Explanation with ABELE
A dedicated interface for an explainer, based on ABELE, for a black-box to classify instances of skin lesions images. The interface is developed to help physicians in the diagnosis of skin cancer. Following the principles of using multiple explanation methods, after classifying an instance, users are presented with two different explanation methods. A counterexample that shows an image classified differently, and a set of exemplar images with the same classification.
ReasonX
REASONX offers declarative, interactive explanations for decision trees, which can be either the primary ML models or global/local surrogate models of any black-box model, making it model-agnostic. It can also integrate background or common sense knowledge through linear constraints. Explanations are presented as factual and contrastive rules, as well as the closest contrastive examples optimized via MILP.
Dr Xai
Doctor XAI offers explanations for predicting a patient's next most probable diagnoses based on their recent clinical history. We developed a visual interface utilizing progressive disclosure to present information related to a specific instance being classified and explained.