A parameterized UCBVI-style algorithm yields distributional regret bounds uniform in δ that achieve optimal expected-distributional trade-offs and confirm a prior conjecture for bandits.
Is reinforcement learning more difficult than bandits? a near-optimal algorithm escaping the curse of horizon
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Unified Framework of Distributional Regret in Multi-Armed Bandits and Reinforcement Learning
A parameterized UCBVI-style algorithm yields distributional regret bounds uniform in δ that achieve optimal expected-distributional trade-offs and confirm a prior conjecture for bandits.