Reproduce “Context-Alignment: Activating and Enhancing LLM Capabilities in Time Series” 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
Yuxiao Hu, Qian Li, Dongxiao Zhang, Jinyue Yan, Yuntian Chen
The authors introduce FSCA, a time-series forecasting method that aligns series and prompts using dual-scale graph neural networks. It represents both token-level details and chunk-level structure, with directed links that connect earlier chunks to prompts and later prediction targets. In zero-shot transfers across eight ETT dataset pairs, FSCA reports average MSE 0.357 versus 0.412 for PatchTST, within this evaluation setup.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.