Multi-Armed Bandit
A classical problem in reinforcement learning and optimization that studies how to optimally allocate a limited trial budget between exploring unknown options and exploiting options known to yield high rewards.
Core Facts
Timeline (last 90 days)
The multi-armed bandit (MAB) is a classic framework in reinforcement learning and online decision-making, in which a player faces K arms, each arm's reward follows an unknown distribution, and the goal is to minimize cumulative regret.
The lower bound for the standard multi-armed bandit problem is O(√T), and ORUCB's extra log(T+1) factor stems from the complexity introduced by simultaneously learning the correction function.
Thompson Sampling models the posterior reward probability of each arm in the multi-armed bandit problem using a Beta distribution, samples from the posterior each round and selects the arm with the highest sampled value, and achieves a theoretically better regret bound than ε-greedy.
In the multi-armed bandit problem, Thompson Sampling each round samples a value from the Beta posterior distribution of each arm, selects the arm with the largest sampled value to execute, and thereby achieves a balance between exploration and exploitation.
All Facts (5)
The core challenge of reinforcement learning lies in the exploration-exploitation tradeoff, where agents need to find a balance between exploiting known effective strategies and exploring unknown possibilities.
75%UnverifiedThe multi-armed bandit (MAB) is a classic framework in reinforcement learning and online decision-making, in which a player faces K arms, each arm's reward follows an unknown distribution, and the goal is to minimize cumulative regret.
50%UnverifiedThe lower bound for the standard multi-armed bandit problem is O(√T), and ORUCB's extra log(T+1) factor stems from the complexity introduced by simultaneously learning the correction function.
50%UnverifiedThompson Sampling models the posterior reward probability of each arm in the multi-armed bandit problem using a Beta distribution, samples from the posterior each round and selects the arm with the highest sampled value, and achieves a theoretically better regret bound than ε-greedy.
50%UnverifiedIn the multi-armed bandit problem, Thompson Sampling each round samples a value from the Beta posterior distribution of each arm, selects the arm with the largest sampled value to execute, and thereby achieves a balance between exploration and exploitation.
50%