An optimistic confidence-interval ranking procedure for best-arm identification across multiple independent bandits yields lower average simple regret and error probability than prior methods when selecting high-performing agents for each game in GVGAI and Ludii.
Bandit-based planning and learning in continuous-action markov decision processes
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Best Agent Identification for General Game Playing
An optimistic confidence-interval ranking procedure for best-arm identification across multiple independent bandits yields lower average simple regret and error probability than prior methods when selecting high-performing agents for each game in GVGAI and Ludii.