Reproduce “Evaluating Time-Series Foundation Models and Multimodal Dietary Context for CGM Forecasting” 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
Bowen Zhang, Hsiu-Wen Cheng, Hongyu Yang, Evie L. Shen, Joleen Vansomphone, Yuna Li, Kerry Zhou, Zitian Qu and 4 more
The paper tests pretrained time-series forecasting for continuous glucose monitoring, comparing zero-shot models with CGM-specific fine-tuning across eight public datasets. Chronos-Bolt-Tiny’s T1D RMSE for a 30-minute forecast after four hours of history falls from 12.21 zero-shot to 10.11 fine-tuned. In a 44-participant meal experiment, nutrients lower postprandial RMSE, but results do not establish clinical benefit or broad generalization.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.