mfederici/dsit

Implementation of the models and datasets used in "An Information-theoretic Approach to Distribution Shifts"

27
/ 100
Experimental

This project provides an implementation of various models and datasets designed to tackle challenges related to 'distribution shifts' in machine learning. It helps researchers and machine learning practitioners evaluate how well their models generalize when the data used for training differs from the data encountered during deployment. You feed in custom datasets, or use provided variants of CMNIST, and it outputs performance metrics showing robustness to these shifts.

No commits in the last 6 months.

Use this if you are a machine learning researcher or practitioner interested in understanding and mitigating distribution shifts, and you need a framework to experiment with models like VIB, DANN, IRM, VREx, or CDANN.

Not ideal if you are looking for an out-of-the-box solution to apply to a specific business problem without deep engagement in machine learning research or model customization.

machine-learning-research model-generalization out-of-domain-prediction robust-ai dataset-analysis
Stale 6m No Package No Dependents
Maintenance 0 / 25
Adoption 7 / 25
Maturity 16 / 25
Community 4 / 25

How are scores calculated?

Stars

25

Forks

1

Language

Jupyter Notebook

License

MIT

Last pushed

Nov 02, 2021

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/mfederici/dsit"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.