Reproduce “Control Foundation Models across Changing Block Diagrams: Generalizing Time-Series Prediction to Dynamic Graphs” 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
Yuji Okamoto, Kenta Hoshino, Masaki Inoue
The control foundation model (CFM) predicts outputs or control inputs from signal histories, masks, and time-varying block-diagram connections. Initialized with pretrained Chronos-2 weights, it uses temporal and graph-based attention, with stage diagrams and switching time supplied. On one simulated unstable linear plant, it produced a decaying closed-loop response and lower input-prediction error than fine-tuned Chronos-2, without establishing generalization.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.