DiCE¶

A library to provide ocunterfactual explanations to ML models – explanations that show how the other outcome (in a binary classifier) can be changed with minimal feature changes. These explainations are often easier for end users to understand (e.g. “you would have obtained the loan if your income was $10,000 higher”).

Logo
Website	http://interpret.ml/DiCE/
Repository	https://github.com/interpretml/DiCE
Byline	Generate Diverse Counterfactual Explanations for any machine learning model.
License	MIT
Project age	3 years 7 months
Backers	Microsoft Researcch (Creator and maintainer)
Lastest News (2022-10-19)	v0.9 Unified API for deep learning and sklearn models; Support for generating CFs without training data (private data mode) for sklearn … more
Size score (1 to 10, higher is better)	2.0
Trend score (1 to 10, higher is better)	7.5

Education Resources¶

URL	Resource Type	Description
https://www.microsoft.com/en-us/research/blog/open-source-library-provides-explanation-for-machine-learning-through-diverse-counterfactuals/	Blog	Blog post on Microsoft Research blog describing DiCE (Jan 2020)
https://arxiv.org/abs/1905.07697	Paper	Paper from FAT 2020 (Conference on Fairness, Accountability, and Transparency)

Git Commit Statistics¶

Statistics computed using Git data through November 30, 2022.

Statistic	Lifetime	Last 12 Months
Commits	589	109
Lines committed	636,041	31,179
Unique committers	18	8
Core committers	6	3

../_images/interpretml_dice-monthly-commits.png

Similar Projects¶

Project	Size Score	Trend Score	Byline
Alibi Explain	3.75	8.25	Algorithms for monitoring and explaining machine learning models.
SHAP	4.75	5.75	A game theoretic approach to explain the output of any machine learning model.
Shapash	4.0	9.0	Shapash makes Machine Learning models transparent and understandable by everyone