Timezone: »

 
Spotlight
Improved Regret Bounds of Bilinear Bandits using Action Space Analysis
Kyoungseok Jang · Kwang-Sung Jun · Se-Young Yun · Wanmo Kang

Wed Jul 21 06:30 PM -- 06:35 PM (PDT) @ None
in Bandits 3 »
We consider the bilinear bandit problem where the learner chooses a pair of arms, each from two different action spaces of dimension $d_1$ and $d_2$, respectively. The learner then receives a reward whose expectation is a bilinear function of the two chosen arms with an unknown matrix parameter $\Theta^*\in\mathbb{R}^{d_1 \times d_2}$ with rank $r$. Despite abundant applications such as drug discovery, the optimal regret rate is unknown for this problem, though it was conjectured to be $\tilde O(\sqrt{d_1d_2(d_1+d_2)r T})$ by Jun et al. (2019) where $\tilde O$ ignores polylogarithmic factors in $T$. In this paper, we make progress towards closing the gap between the upper and lower bound on the optimal regret. First, we reject the conjecture above by proposing algorithms that achieve the regret $\tilde O(\sqrt{d_1 d_2 (d_1+d_2) T})$ using the fact that the action space dimension $O(d_1+d_2)$ is significantly lower than the matrix parameter dimension $O(d_1 d_2)$. Second, we additionally devise an algorithm with better empirical performance than previous algorithms.
Keywords: Bandits · Reinforcement Learning and Planning
[ Paper ]

Author Information

Kyoungseok Jang (KAIST)
Kwang-Sung Jun (University of Arizona)
Se-Young Yun (KAIST)
Wanmo Kang (KAIST)

Related Events (a corresponding poster, oral, or spotlight)

More from the Same Authors