ResearchBenchmarksRegret Minimization on Simple World CommFor AI assistantsMarkdownContent indexRegret Minimization on Simple World CommFollow17.26Maximum RegretNePPO15.879225.199634.5243.8404Mar 7, 2026Evaluation resultsMethodMethodLinksMaximum RegretNePPO2026.0317.26IPPO2026.0323.9MAPPO2026.0351.78