ResearchBenchmarksPolicy Optimization on Office World MAP1Follow3,125Avg Training StepsQR-MAXRM-49,552.8306,022.35661,597.51,017,172.65Dec 16, 2025Evaluation ResultsMethodMethodLinksAvg Training StepsQR-MAXRMExp=EXP5Exp=EXP52025.123,125QR-MAXExp=EXP5Exp=EXP52025.1224,222QRMExp=EXP5Exp=EXP52025.12225,140UCBVI-sBExp=EXP5Exp=EXP52025.12250,800R-MAXExp=EXP5Exp=EXP52025.12272,080UCBVI-BExp=EXP5Exp=EXP52025.12555,977UCBVI-HExp=EXP5Exp=EXP52025.121,320,070