Parameter-Efficient Active Learning for Foundational models

TL;DR

This paper explores combining parameter-efficient fine-tuning with active learning to improve data sampling in vision transformer models, especially for out-of-distribution image datasets under strict budget constraints.

Contribution

It introduces a novel approach integrating parameter-efficient fine-tuning with active learning for foundational models, enhancing sampling efficiency in challenging vision tasks.

Findings

Improved active learning performance on out-of-distribution datasets

Enhanced data annotation efficiency with parameter-efficient methods

Demonstrated strategic advantages in budget-constrained scenarios

Abstract

Foundational vision transformer models have shown impressive few shot performance on many vision tasks. This research presents a novel investigation into the application of parameter efficient fine-tuning methods within an active learning (AL) framework, to advance the sampling selection process in extremely budget constrained classification tasks. The focus on image datasets, known for their out-of-distribution characteristics, adds a layer of complexity and relevance to our study. Through a detailed evaluation, we illustrate the improved AL performance on these challenging datasets, highlighting the strategic advantage of merging parameter efficient fine tuning methods with foundation models. This contributes to the broader discourse on optimizing AL strategies, presenting a promising avenue for future exploration in leveraging foundation models for efficient and effective data annotation in specialized domains.