2025
Exploration via Feature Perturbation in Contextual Bandits
NeurIPS 2025spotlight
We propose *feature perturbation*, a simple yet effective exploration strategy for contextual bandits that injects randomness directly into feature inputs, instead of randomizing unknown parameters or adding noise to rewards. Remarkably, this algorithm achieves $\widetilde{\mathcal{O}}(d\sqrt{T})$ w…