The Performance of Transferability Metrics does not Translate to Medical Tasks

TL;DR

This paper evaluates the effectiveness of transferability metrics in medical image analysis, revealing that they do not reliably predict performance in medical tasks despite success in general datasets.

Contribution

It provides a comprehensive evaluation of seven transferability scores on medical datasets, highlighting their limitations and the need for specialized metrics in medical applications.

Findings

Transferability scores do not reliably estimate performance in medical datasets.

Current transferability metrics fail in out-of-distribution medical scenarios.

Further research is needed to develop better transferability measures for medical tasks.

Abstract

Transfer learning boosts the performance of medical image analysis by enabling deep learning (DL) on small datasets through the knowledge acquired from large ones. As the number of DL architectures explodes, exhaustively attempting all candidates becomes unfeasible, motivating cheaper alternatives for choosing them. Transferability scoring methods emerge as an enticing solution, allowing to efficiently calculate a score that correlates with the architecture accuracy on any target dataset. However, since transferability scores have not been evaluated on medical datasets, their use in this context remains uncertain, preventing them from benefiting practitioners. We fill that gap in this work, thoroughly evaluating seven transferability scores in three medical applications, including out-of-distribution scenarios. Despite promising results in general-purpose datasets, our results show that no transferability score can reliably and consistently estimate target performance in medical contexts, inviting further work in that direction.