Thompson Sampling for Adversarial Bit Prediction
This provides theoretical insights for online learning algorithms in adversarial environments, but it is incremental as it focuses on specific regret characterizations.
The paper tackles the problem of analyzing Thompson sampling in adversarial bit prediction, characterizing sequences with smallest and largest expected regret and bounding regret as O(√T) for worst-case and O(1) for best-case sequences, with extensions to weighted error models.
We study the Thompson sampling algorithm in an adversarial setting, specifically, for adversarial bit prediction. We characterize the bit sequences with the smallest and largest expected regret. Among sequences of length $T$ with $k < \frac{T}{2}$ zeros, the sequences of largest regret consist of alternating zeros and ones followed by the remaining ones, and the sequence of smallest regret consists of ones followed by zeros. We also bound the regret of those sequences, the worse case sequences have regret $O(\sqrt{T})$ and the best case sequence have regret $O(1)$. We extend our results to a model where false positive and false negative errors have different weights. We characterize the sequences with largest expected regret in this generalized setting, and derive their regret bounds. We also show that there are sequences with $O(1)$ regret.