Reproduce “Tactile-JEPA: Topology-Aware Self-Supervised Representation Learning for Distributed Tactile Sensors” 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

Elizaveta Kovtun, Matvey Konovalov, Andrey Sakhovskiy, Semen Budennyy

Tactile-JEPA learns representations for distributed tactile skins by masking taxels according to their connectivity graph and predicting hidden taxel embeddings from visible signals. It mixes connected local and scattered global masks instead of reconstructing raw readings. On Sparsh-skin, force RMSE is 14.05 ± 0.73 cN and orientation θ-RMSE 5.48 ± 0.39°, with transfer across tested datasets but not best policy performance.

Summary by alphaXiv.

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