← 研踪 PaperTrailarXiv:2611.23801 · v1
Improved Leverage Score Sampling for Constrained Active Linear Regression
首次提交 2026-11-18 12:06
版本更新 2026-11-18 12:06
雷达观测 2026-11-19 19:08
摘要
In this paper we study active learning for constrained linear regression. Given a design matrix , a constraint set , and query access to , we seek an approximate solution to using as few queries as possible. This problem arises in many scientific and engineering applications, and randomized leverage score sampling provides a powerful method to solve it. However, existing leverage score methods ignore the constraint or encode it through ridge regularization, suffering unfavorable query complexity for anisotropic constraint sets. We propose ellipsoid-regularized leverage scores, where we approximate the constraint set by an outer ellipsoid and use its shape matrix to bias sampling toward directions that matter under the constraint geometry. We prove that the query complexity is governed by an anisotropic effective dimension . With appropriate parameters, queries suffice to return a point in with objective value at most times the optimum. This bound can substantially improve on the standard leverage-score bound, since anisotropic constraints can make much smaller than . For convex constraints, we improve the dependence on from to . We also prove two lower bounds: first, the dependence is unavoidable without convexity assumptions on , and second, in some regimes, the effective dimension dependence is necessary for ellipsoidal constraints. Experiments on synthetic regression problems and a parametric heat conduction problem show lower regression error than standard baselines for similar query budget.
作者备注:To appear in NeurIPS 2026 (Spotlight)
观测到的版本
相邻记录可能跳过中间版本;首次发现 v2 及以上也会标为版本更新。
v1 · 2026-11-18 12:06 · 最新观测版本
Improved Leverage Score Sampling for Constrained Active Linear Regression
In this paper we study active learning for constrained linear regression. Given a design matrix , a constraint set , and query access to , we seek an approximate solution to using as few queries as possible. This problem arises in many scientific and engineering applications, and randomized leverage score sampling provides a powerful method to solve it. However, existing leverage score methods ignore the constraint or encode it through ridge regularization, suffering unfavorable query complexity for anisotropic constraint sets. We propose ellipsoid-regularized leverage scores, where we approximate the constraint set by an outer ellipsoid and use its shape matrix to bias sampling toward directions that matter under the constraint geometry. We prove that the query complexity is governed by an anisotropic effective dimension . With appropriate parameters, queries suffice to return a point in with objective value at most times the optimum. This bound can substantially improve on the standard leverage-score bound, since anisotropic constraints can make much smaller than . For convex constraints, we improve the dependence on from to . We also prove two lower bounds: first, the dependence is unavoidable without convexity assumptions on , and second, in some regimes, the effective dimension dependence is necessary for ellipsoidal constraints. Experiments on synthetic regression problems and a parametric heat conduction problem show lower regression error than standard baselines for similar query budget.
查看此版本解读 →查看此版本原文 ↗