Towards Estimating Transferability using Hard Subsets

TL;DR

This paper introduces HASTE, a method to estimate transferability of source models to target tasks using hard data subsets, improving the reliability of transferability metrics without expensive fine-tuning.

Contribution

HASTE is a novel approach that leverages hard data subsets and internal model representations to enhance transferability estimation accuracy.

Findings

HASTE improves the reliability of existing transferability metrics.

HASTE performs consistently better or on par with state-of-the-art metrics.

Theoretical bounds are validated empirically across multiple datasets.

Abstract

As transfer learning techniques are increasingly used to transfer knowledge from the source model to the target task, it becomes important to quantify which source models are suitable for a given target task without performing computationally expensive fine tuning. In this work, we propose HASTE (HArd Subset TransfErability), a new strategy to estimate the transferability of a source model to a particular target task using only a harder subset of target data. By leveraging the internal and output representations of model, we introduce two techniques, one class agnostic and another class specific, to identify harder subsets and show that HASTE can be used with any existing transferability metric to improve their reliability. We further analyze the relation between HASTE and the optimal average log likelihood as well as negative conditional entropy and empirically validate our theoretical bounds. Our experimental results across multiple source model architectures, target datasets, and transfer learning tasks show that HASTE modified metrics are consistently better or on par with the state of the art transferability metrics.