Reproduce “A study of animal action segmentation algorithms across supervised, unsupervised, and semi-supervised learning paradigms” on OpenResearch

Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.

Universal app for Apple silicon and Intel. Requires macOS 11 or later.

About this paper

Ari Blau, Evan S Schaffer, Neeli Mishra, Nathaniel J Miska, International Brain Laboratory, Liam Paninski, Matthew R Whiteway

The authors introduce a semi-supervised switching state-space model that learns from a small set of human-labeled frames and unlabeled behavioral data. They compare it with supervised and unsupervised methods across four datasets. Adding motion features improves performance, but the supervised temporal convolutional network aligns best with human labels; this result does not establish superior behavioral discovery.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.