One-shot Active Learning Based on Lewis Weight Sampling for Multiple Deep Models

TL;DR

This paper introduces a one-shot active learning method for multiple deep models that reduces computational costs by avoiding iterative training, using Lewis weight sampling for efficient label query selection.

Contribution

It proposes a novel one-shot active learning approach leveraging Lewis weight sampling and multiple representations, enabling efficient training of multiple models without repeated training cycles.

Findings

Achieves competitive performance on 11 benchmarks.

Reduces computational cost compared to iterative methods.

Provides theoretical bounds on sample complexity.

Abstract

Active learning (AL) for multiple target models aims to reduce labeled data querying while effectively training multiple models concurrently. Existing AL algorithms often rely on iterative model training, which can be computationally expensive, particularly for deep models. In this paper, we propose a one-shot AL method to address this challenge, which performs all label queries without repeated model training. Specifically, we extract different representations of the same dataset using distinct network backbones, and actively learn the linear prediction layer on each representation via an $\ell_p$-regression formulation. The regression problems are solved approximately by sampling and reweighting the unlabeled instances based on their maximum Lewis weights across the representations. An upper bound on the number of samples needed is provided with a rigorous analysis for $p\in [1, +\infty)$. Experimental results on 11 benchmarks show that our one-shot approach achieves competitive performances with the state-of-the-art AL methods for multiple target models.