Reproduce “TimeCap: A Scalable Open-Source Pipeline for Captioning Wearable Health Time-Series Data” 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
Max Rosenblattl, Milana Gurbanova, Karl Aurel Deck, Narayan Schuetz, Thomas Kaar, Manuel Milling, Nicolas Zumarraga, Ning Wang and 11 more
TimeCap converts wearable signals into reproducible text captions using fixed, configurable rules, including descriptions that combine labeled activity intervals with synchronized evidence from other channels. Applied to OpenMHC, it produced 7,005,763 captions across 908,944 person-days and 19 channels. OpenTSLM trained with captions outperformed its label-only counterpart on 29 of 32 comparisons, measuring self-reported attributes rather than diagnosis or clinical reliability.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.