Which Model to Transfer? A Survey on Transferability Estimation

TL;DR

This survey reviews recent advances in transferability estimation for pre-trained models, categorizing methods into source-free and source-dependent, and discusses challenges and future directions in the field.

Contribution

It provides the first comprehensive review and taxonomy of transferability estimation methods, clarifying definitions and experimental settings.

Findings

Categorizes transferability estimation into source-free and source-dependent methods.

Provides a systematic taxonomy and comprehensive overview of existing techniques.

Outlines challenges and future research directions in transferability estimation.

Abstract

Transfer learning methods endeavor to leverage relevant knowledge from existing source pre-trained models or datasets to solve downstream target tasks. With the increase in the scale and quantity of available pre-trained models nowadays, it becomes critical to assess in advance whether they are suitable for a specific target task. Model transferability estimation is an emerging and growing area of interest, aiming to propose a metric to quantify this suitability without training them individually, which is computationally prohibitive. Despite extensive recent advances already devoted to this area, they have custom terminological definitions and experimental settings. In this survey, we present the first review of existing advances in this area and categorize them into two separate realms: source-free model transferability estimation and source-dependent model transferability estimation. Each category is systematically defined, accompanied by a comprehensive taxonomy. Besides, we address challenges and outline future research directions, intending to provide a comprehensive guide to aid researchers and practitioners.