# EthicalML/xai

XAI - An eXplainability toolbox for machine learning

Repository: https://github.com/EthicalML/xai
Canonical: https://ross.abutalabs.com/products/xai
Homepage: https://ethical.institute/principles.html#commitment-3
Language: Python
License: MIT
License Family: permissive
Topics: explainability, xai, ml, ai, bias, artificial-intelligence, bias-evaluation, explainable-ai, explainable-ml, machine-learning, machine-learning-explainability, interpretability, xai-library, evaluation, imbalance, upsampling, downsampling, feature-importance
Last push: 2025-11-29T12:57:47+00:00

## Health v2 (maintenance only)
Score: 65/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 54, release rhythm 59, longevity 100
- inputs: {"age_days": 2791, "days_push": 277, "days_rel": 277, "gap_med": 0, "n_releases_24m": 2}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1259, forks 186 (observed 2026-08-28T04:04:09.885584+00:00)

## What it is
XAI is a Python library from the Institute for Ethical AI & ML that provides tools for analysing and evaluating machine learning datasets and models with explainability in mind. It supports bias evaluation, class imbalance handling (upsampling/downsampling), and feature importance analysis across the three steps of explainable ML: identifying, mitigating, and evaluating.

## Use cases
- evaluate bias in machine learning datasets
- measure feature importance of trained models
- handle class imbalance with upsampling or downsampling
- explain model predictions for fairness audits
- analyse datasets for responsible ML practices
- identify discrepancies causing sub-optimal model performance

## When to choose
- you need a lightweight Python toolbox for bias and imbalance analysis
- you want explainability tooling aligned with responsible ML principles
- you are doing exploratory fairness evaluation in Jupyter notebooks

## When to avoid
- you need production-grade, actively maintained explainability tooling
- you need model-agnostic explainers like SHAP or LIME with broad model support
- you require support for modern Python versions beyond 3.7

## Facets
- artifact type: library
- maturity: experimental
- function: machine-learning, data-science, data-visualization, benchmarking
- domain: machine-learning, artificial-intelligence, data-science
- platform: python
- tags: explainable-ai, xai, bias-evaluation, interpretability, feature-importance, imbalance, fairness, algorithms

## Member repositories
- EthicalML/xai (main) score 65

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:04:09.885584+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T05:05:27.885818+00:00, confidence not recorded.
  - readme: https://github.com/EthicalML/xai (fetched 2026-08-28T04:04:09.885584+00:00, sha 35b861bf83d2)
  - homepage: https://ethical.institute/principles.html#commitment-3 (fetched 2026-08-29T12:17:16.958713+00:00, sha c102e7fa21cf)
  - registry_pypi: https://pypi.org/pypi/xai/json (fetched 2026-08-29T12:17:16.961182+00:00, sha 43f893701a86)
- Data as of 2026-08-30T08:39:29.467469+00:00.