cs.LGJun 10, 2026

Efficient Multinomial Logistic Bandit via Frequent Directions

Authors: Linzhe HeYu-Jie ZhangSifan YangLijun Zhang

Abstract

This paper studies efficient online algorithms for multinomial logistic bandits (MLogB), where the feedback distribution over K+1K+1 outcomes follows a multinomial logistic model of dd-dimensional action vectors. A representative UCB-type algorithm, OFUL-MLogB, achieves a regret bound of O~(KdT)\tilde{\mathcal{O}}(Kd\sqrt{T}), but still requires O(K3d3)\mathcal{O}(K^3d^3) time and O(K2d2)\mathcal{O}(K^2d^2) space per round due to parameter estimation and optimistic reward construction, which is prohibitive in high-dimensional settings. To address this limitation, we propose EOFD-MLogB, which integrates frequent directions matrix sketching into OFUL-MLogB. By maintaining a low-rank SVD sketch of the accumulated Hessian, constrained online Newton updates in parameter estimation and Kd×KKd \times K spectral-norm computations in the reward bonus are reduced to one-dimensional root-finding tasks and K×KK \times K eigenvalue computations, respectively. This yields dominant per-round time complexity O(Kd(m+K)2)\mathcal{O}(Kd(m+K)^2) and space complexity O(Kd(m+K))\mathcal{O}(Kd(m+K)), where mdm \ll d is the sketch size. We further prove a regret bound of O~(ΔT(KdlnΔT+m)T)\tilde{\mathcal{O}}(Δ_T(Kd\lnΔ_T+m)\sqrt{T}), where the sketching error factor ΔTΔ_T is controlled by the mm-truncated spectral tail of the Hessian. Thus, when the Hessian is approximately low-rank, the regret is close to that of OFUL-MLogB. Experiments validate the computational efficiency and competitive performance.

Explore similar work

CardsList