Eluder dimension: localise it!
Alireza Bakhtiari ⋅ Alex Ayoub ⋅ Samuel Robertson ⋅ David Janz ⋅ Csaba Szepesvari
2025 Spotlight Poster
Abstract
We establish a lower bound on the eluder dimension in generalised linear model classes, showing that standard eluder dimension-based analysis cannot lead to first-order regret bounds. To address this, we introduce a localisation method for the eluder dimension; our analysis immediately recovers and improves on classic results for Bernoulli bandits, and allows for the first genuine first-order bounds for finite-horizon reinforcement learning tasks with bounded cumulative returns.
Video
Chat is not available.
Successful Page Load