We study adversarial combinatorial bandits with $m$-set actions, where at each round the learner selects $m$ out of $d$ items and observes only the aggregate loss of the selected items. The resulting action set contains $K=\binom{d}{m}$ elements and can therefore be exponentially large. Nevertheless, the loss of every action is determined by the same $d$-dimensional vector of item losses. We propose a computationally efficient algorithm that exploits this structure without explicitly enumerating the action set. Against adaptive non-anticipating adversaries, it guarantees, with probability at least $1-\delta$, regret against the best fixed action of \[ R_T = O\left(\sqrt{dT\log(K/\delta)}\right). \] This matches the high-probability regret bound of the finite-action EXP3-KW algorithm of Zimmert and Lattimore, whose direct implementation may require exponential space. Our algorithm instead represents each sampling distribution with $d$ parameters and runs in polynomial time without enumerating the action set. Thus, it resolves the open problem posed by Maiti et al.
An algorithm is designed that achieves second-order path-length regret in adversarial $K$-armed bandits against oblivious loss sequences using an adaptive restart scheme whose path-length estimator has uniformly bounded increments.
Let $p_0,p_1,\ldots,p_N$ be points of the unit ball of $\mathbb R^d$, processed in a prescribed order. We study the insertion cost $\sum_{i=1}^N\lVert p_i-x_{i-1}\rVert^\alpha$, where each $x_{i-1}$ is computed from the previously observed points. The input-order path is sensitive to the input distribution but can repe...
An adaptive adversary is constructed that refines a hidden large-reward subspace based on the learner's actions, in such a way that low regret is impossible without estimating the subspace; as a result, lower-bounding the regret reduces to studying the arising subspace estimation problem.
Moise Blanchard, Dmitrii M. Ostrovskii, Aadirupa Saha· arXiv.org· 0 citations
An expected approximation guarantee of an expected approximation guarantee of $8(\ln k+2)\left(\frac{1+\varepsilon}{1-\varepsilon}{1-\varepsilon}\right)^4 = (1+O(\varepsilon))\,8(\ln k+2)$.
The proofs combine a persistent randomized Balog-Szemer\'edi-Gowers procedure producing a fixed implicit small-doubling subset on the $\sqrt{\alpha}$ retained-mass scale, conditionally exact finite product sampling, size-oblivious algorithmic PFR, and deterministic lifting.
We study expected improvement (EI) for minimizing a deterministic function $f$ in the RKHS $\mathcal H_k$ of a continuous positive-semidefinite kernel $k$ on a nonempty compact set $\mathcal X\subset\mathbb R^d$. Function values are observed exactly, and EI is computed from a fixed zero-mean Gaussian-process model with...
Emmanuel Vazquez, S. Petit· arXiv.org· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
