TimeRSI: Self-improving agent system for time-series forecasting models
Overview
Deqing Fu announces TimeRSI, a self-improving agentic deployment system for time-series foundation models.
Fu says that with TimeRSI, TimesFM-3.0 returns to No. 1 on the TIME and FEV-bench leaderboards as of October 15. Fu also says TabFM-Auto, a coding agent that evolves the pipeline around a frozen tabular foundation model, beats the field. These are Fu's claims in a launch post on X; no independent leaderboard verification is reported.
Written by AI from the articles below · updated Oct 15, 9:04 PM ET
Check the sources:
Article timeline
The articles in this story. Times are ET.
Deqing Fu@DeqingFuXTimeRSI, a self-improving agentic harness, returns TimesFM-3.0 to No. 1 on TIME and FEV-benchAIIntroducing TimeRSI, a self-improving agentic deployment system for time-series foundation models, Deqing Fu says. With TimeRSI, TimesFM-3.0 returns to No. 1 on the TIME and FEV-bench leaderboards as of October 15. Fu also says TabFM-Auto, a coding agent that evolves the pipeline around a frozen tabular foundation model, beats the field.

Heat trend
Not enough continuous observations to show a trend yet.