lapalap/dora

GitHub repository for DORA: Data-agnOstic Representation Analysis paper. DORA allows to find outlier representations in Deep Neural Networks.

33
/ 100
Emerging

This tool helps machine learning engineers and researchers analyze Deep Neural Networks for unintended learning. It takes your pre-trained neural network as input and outputs identified 'infected' neurons or representations that might be learning spurious, irrelevant concepts. The user is a machine learning practitioner who wants to ensure their models are robust and reliable.

No commits in the last 6 months.

Use this if you need to automatically detect and understand anomalous behaviors or 'infected' components within your deep learning models.

Not ideal if you are looking for a tool to explain overall model predictions to a non-technical audience or to interpret why a specific input led to a specific output.

deep-learning-diagnostics model-robustness ai-safety neural-network-analysis representation-learning
Stale 6m No Package No Dependents
Maintenance 0 / 25
Adoption 7 / 25
Maturity 16 / 25
Community 10 / 25

How are scores calculated?

Stars

27

Forks

3

Language

Python

License

MIT

Last pushed

Mar 19, 2023

Commits (30d)

0

Get this data via API

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

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