Regret Minimization on Multi-armed Bandit anytime setting
2Regret CoefficientNaive implementation
Evaluation Results
| Method | Links | |
|---|---|---|
| Naive implementationTime Complexity per Round=O(K) worst-case2025.12 | 2 | |
| Binary tree EXP3 (Section 3)Time Complexity per Round=O(log K) worst-case2025.12 | 2 | |
| Constant-Time EXP3 / Advanced data structure (Section 4)Time Complexity per Round=O(1) expected2025.12 | 2 | |
| Alias method EXP3 (Section 5)Time Complexity per Round=O(1) expected2025.12 | 2 | |
| Chewi et al.Time Complexity per Round=O(log^2 K) worst-case processing + O(log K) expected sampling2025.12 | 4 |