AI RESEARCH
Hedging Memory Horizons for Non-Stationary Prediction via Online Aggregation
arXiv CS.LG
•
ArXi:2605.06541v1 Announce Type: new We study online prediction under distribution shift, where inputs arrive chronologically and outcomes are revealed only after prediction. In this setting, predictors must remain stable in quiet regimes yet adapt when regimes shift, and the right adaptation memory is unknown in advance.