Convexity of representations
This package contains methods to compute convexity scores to measure convexity of latent representations of neural networks as defined in
Tětková, L., Brüsch, T., Dorszewski, T. et al. On convex decision regions in deep network representations. Nat Commun 16, 5419 (2025). https://doi.org/10.1038/s41467-025-60809-y
Paper: https://www.nature.com/articles/s41467-025-60809-y
Documentation: https://nnconvexity.readthedocs.io/en/latest/
Source code: https://github.com/LenkaTetkova/nnconvexity
See code for the paper containing also a demo for using this package.
It provides functions for two types of convexity:
- Euclidean: sample points on a segment between two points of the same class and evaluate whether they are classified into the same class.
- graph: approximation of convexity on a manifold -- construct a graph based on nearest neighbors and evaluate proportion of the shortest paths that go through the same class.
Metadata
Release files for nnconvexity 0.1.1
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | Provenance | Actions |
|---|---|---|---|---|
| nnconvexity-0.1.1.tar.gz | 2.8 MB | Without provenance | Details for nnconvexity-0.1.1.tar.gz |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Actions Reset |
|---|---|---|---|---|
| nnconvexity-0.1.1-py3-none-any.whl | Python 3 | none | any | Details for nnconvexity-0.1.1-py3-none-any.whl |
Total release size: 2.8 MB
Release files / nnconvexity-0.1.1.tar.gz
| Download URL | nnconvexity-0.1.1.tar.gz |
|---|---|
| Size | 2.8 MB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
91d9c9df0bdf00fa80668e1dac43ea412f6e9c3d5ab797418a20371dcc6d8db6
|
|
BLAKE2b-256 checksum How to use checksums |
aca2d7ee10fd2d48e3787d684973d54e3bfe73f841ded5d9d88bcb57bb2b3955
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
uv/0.7.19
|
Release files / nnconvexity-0.1.1-py3-none-any.whl
| Download URL | nnconvexity-0.1.1-py3-none-any.whl |
|---|---|
| Size | 6.0 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
52bdf8105923724bfe9b65bf439092eb2c7b3d9363d83fac8171f0a6eb630ce8
|
|
BLAKE2b-256 checksum How to use checksums |
12483fdaa33f5168da8b567f90ce4108b826e273a40e049634d9c37d686da18c
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
uv/0.7.19
|
Security for nnconvexity 0.1.1
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
No provenance found
None of the files in this release were uploaded with provenance metadata.
A note on package security
PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.
To learn more about protecting yourself from malicious actors, visit our package security documentation.