K-UCBVI learns from K-ary episodic human feedback via maximum-likelihood reward estimation plus optimism, with a proven O(sqrt(N) log N) regret bound.
Proof of Theorem 3 The optimistic reward function is defined as follows: R(bwn,τ ) = min R(bwn,τ ) + 4Kexp(4B) ηλmin(ΣDn) r C2 2n log 4 δ, K− 1
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
REJECT 1representative citing papers
citing papers explorer
-
Reinforcement Learning from Multi-level and Episodic Human Feedback
K-UCBVI learns from K-ary episodic human feedback via maximum-likelihood reward estimation plus optimism, with a proven O(sqrt(N) log N) regret bound.