Bandit Guided Submodular Curriculum for Adaptive Subset Selection

Prateek Chanda, Prayas Agrawal, Saral Sureka, Lokesh Reddy Polu, Atharv Kshirsagar, Ganesh Ramakrishnan

Keywords: Curriculum Learning, Adaptive Subset Selection, Multi Arm Bandit, Submodular Functions, No-regret Analysis

TL;DR: We propose an adaptive subset selection framework for supervised learning, where submodular functions quantify sample difficulty to guide curriculum learning.

Abstract: Traditional curriculum learning proceeds from easy to hard samples, yet defining a reliable notion of difficulty remains elusive. Prior work has used submodular functions to induce difficulty scores in curriculum learning. We reinterpret adaptive subset selection and formulate it as a multi-armed bandit problem, where each arm corresponds to a submodular function guiding sample selection. We introduce OnlineSubmod, a novel online greedy policy that optimizes a utility-driven reward and provably achieves no-regret performance under various sampling regimes. Empirically, OnlineSubmod outperforms both traditional curriculum learning and bi-level optimization approaches across vision and language datasets, showing superior accuracy-efficiency tradeoffs. More broadly, we show that validation-driven reward metrics offer a principled way to guide the curriculum schedule. Our code is publicly available at GitHub : https://github.com/efficiency-learning/banditsubmod/.

Primary Area: General machine learning (supervised, unsupervised, online, active, etc.)

Submission Number: 28216