paper-with-me

홈 › Papers

An Empirical Process Approach to the Union Bound: Practical Algorithms for Combinatorial and Linear Bandits

2020-06-21 · NeurIPS 2020 12 · Julian Katz-Samuels, Lalit Jain, Zohar Karnin, Kevin Jamieson

This paper proposes near-optimal algorithms for the pure-exploration linear bandit problem in the fixed confidence and fixed budget settings. Leveraging ideas from the theory of suprema of empirical processes, we provide an algorithm whose sample complexity scales with the geometry of the instance and avoids an explicit union bound over the number of arms. Unlike previous approaches which sample based on minimizing a worst-case variance (e.g. G-optimal design), we define an experimental design objective based on the Gaussian-width of the underlying arm set. We provide a novel lower bound in terms of this objective that highlights its fundamental role in the sample complexity. The sample complexity of our fixed confidence algorithm matches this lower bound, and in addition is computationally efficient for combinatorial classes, e.g. shortest-path, matchings and matroids, where the arm sets can be exponentially large in the dimension. Finally, we propose the first algorithm for linear bandits in the the fixed budget setting. Its guarantee matches our lower bound up to logarithmic factors.

📄 PDF Abstract BibTeX arXiv:2006.11685

Code (0)

등록된 구현이 없습니다.

Tasks

Experimental Design

Similar Papers 제목 키워드 기반

A note on the smallest eigenvalue of the empirical covariance of causal Gaussian processes

2022-12-19 · Ingvar Ziemann

We present a simple proof for bounding the smallest eigenvalue of the empirical covariance in a causal Gaussian process. Along the way, we establish a one-sided tail inequality for Gaussian quadratic forms using a causal…

Gaussian Processes

A PAC-Bayesian View of Generalisation for Physics-Informed Machine Learning

2026-05-25 · Thien V. Nguyen, Amaury Habrard, Benjamin Guedj arxiv

Physics-informed machine learning (PIML) integrates mechanistic knowledge, typically in the form of partial differential equations (PDE), into data-driven models. Despite strong empirical performance, its statistical gen…

Online Heavy-tailed Change-point detection

2023-06-15 · Abishek Sankararaman, Balakrishnan, Narayanaswamy

We study algorithms for online change-point detection (OCPD), where samples that are potentially heavy-tailed, are presented one at a time and a change in the underlying mean must be detected as early as possible. We pre…

Change Point Detection

Model Selection for Nonnegative Matrix Factorization by Support Union Recovery

2018-10-23 · Zhaoqiang Liu

Nonnegative matrix factorization (NMF) has been widely used in machine learning and signal processing because of its non-subtractive, part-based property which enhances interpretability. It is often assumed that the late…

Model Selection

Auditing of Unlearning Algorithms

2026-07-07 · Sahasrajit Sarmasarkar, Anastasia Koloskova, Sanmi Koyejo arxiv

Evaluating whether unlearning algorithms truly remove training data influence remains an open challenge. We propose a practical auditor that computes data-dependent lower bounds on the unlearning parameter $\varepsilon$ …