Improved Online Confidence Bounds for Multinomial Logistic Bandits

In this paper, we propose an improved online confidence bound for multinomial logistic (MNL) models and apply this result to MNL bandits, achieving variance-dependent optimal regret. Recently, Lee & Oh (2024) established an online confidence bound for MNL models and achieved nearly minimax-optimal regret in MNL bandits. However, their results still depend on the normboundedness of the unknown parameter B and the maximum size of possible outcomes K. To address this, we first derive an online confidence bound of O ´?d log t `B? d ¯, which is a significant improvement over the previous bound of OpB ? d log t log Kq (Lee & Oh, 2024). This is mainly achieved by establishing tighter selfconcordant properties of the MNL loss and applying Ville's inequality to bound the estimation error. Using this new online confidence bound, we propose a constant-time algorithm, OFU-MNL++, which achieves a variance-dependent regret bound of O ´d log T b ř T t"1 σ 2 t ¯for sufficiently large T , where σ 2 t denotes the variance of the rewards at round t, d is the dimension of the contexts, and T is the total number of rounds. Furthermore, we introduce a Maximum Likelihood Estimation (MLE)-based algorithm, OFU-M 2 NL, which achieves an anytime polypBq-free regret of O ´d logpBT q b ř T t"1 σ 2 t ¯.