Understanding the Transferability of Representations via Task-Relatedness

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Mehra, Akshay, Zhang, Yunbei, Hamm, Jihun
Format: Preprint
Published: 2023
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910673610473472
author Mehra, Akshay
Zhang, Yunbei
Hamm, Jihun
author_facet Mehra, Akshay
Zhang, Yunbei
Hamm, Jihun
contents The growing popularity of transfer learning, due to the availability of models pre-trained on vast amounts of data, makes it imperative to understand when the knowledge of these pre-trained models can be transferred to obtain high-performing models on downstream target tasks. However, the exact conditions under which transfer learning succeeds in a cross-domain cross-task setting are still poorly understood. To bridge this gap, we propose a novel analysis that analyzes the transferability of the representations of pre-trained models to downstream tasks in terms of their relatedness to a given reference task. Our analysis leads to an upper bound on transferability in terms of task-relatedness, quantified using the difference between the class priors, label sets, and features of the two tasks. Our experiments using state-of-the-art pre-trained models show the effectiveness of task-relatedness in explaining transferability on various vision and language tasks. The efficient computability of task-relatedness even without labels of the target task and its high correlation with the model's accuracy after end-to-end fine-tuning on the target task makes it a useful metric for transferability estimation. Our empirical results of using task-relatedness to select the best pre-trained model from a model zoo for a target task highlight its utility for practical problems.
format Preprint
id arxiv_https___arxiv_org_abs_2307_00823
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle Understanding the Transferability of Representations via Task-Relatedness
Mehra, Akshay
Zhang, Yunbei
Hamm, Jihun
Machine Learning
The growing popularity of transfer learning, due to the availability of models pre-trained on vast amounts of data, makes it imperative to understand when the knowledge of these pre-trained models can be transferred to obtain high-performing models on downstream target tasks. However, the exact conditions under which transfer learning succeeds in a cross-domain cross-task setting are still poorly understood. To bridge this gap, we propose a novel analysis that analyzes the transferability of the representations of pre-trained models to downstream tasks in terms of their relatedness to a given reference task. Our analysis leads to an upper bound on transferability in terms of task-relatedness, quantified using the difference between the class priors, label sets, and features of the two tasks. Our experiments using state-of-the-art pre-trained models show the effectiveness of task-relatedness in explaining transferability on various vision and language tasks. The efficient computability of task-relatedness even without labels of the target task and its high correlation with the model's accuracy after end-to-end fine-tuning on the target task makes it a useful metric for transferability estimation. Our empirical results of using task-relatedness to select the best pre-trained model from a model zoo for a target task highlight its utility for practical problems.
title Understanding the Transferability of Representations via Task-Relatedness
topic Machine Learning
url https://arxiv.org/abs/2307.00823