AI RESEARCH
Cascading Bandits With Feedback
arXiv CS.LG
•
ArXi:2511.10938v2 Announce Type: replace Motivated by the challenges of edge inference, we study a variant of the cascade bandit model in which each arm corresponds to an inference model with an associated accuracy and error probability. We analyse four decision-making policies-Explore-then-Commit, Action Elimination, Lower Confidence Bound (LCB), and Thompson Sampling-and provide sharp theoretical regret guarantees for each.