Multi-ArmedBanditswithBoundedArm-Memory: Near-OptimalGuaranteesforBest-Arm IdentificationandRegretMinimization
–Neural Information Processing Systems
In this setting, the arms arrive in a stream, and the number of arms that can be storedinthememory atanytime,isbounded.
Neural Information Processing Systems
Feb-19-2026, 06:54:25 GMT
- Technology: