Has Your Pretrained Model Improved? A Multi-head Posterior Based Approach

Fuente: arXiv
Enregistré dans:
Détails bibliographiques
Auteurs principaux: Aboagye, Prince, Zheng, Yan, Wang, Junpeng, Saini, Uday Singh, Dai, Xin, Yeh, Michael, Fan, Yujie, Zhuang, Zhongfang, Jain, Shubham, Wang, Liang, Zhang, Wei
Format: Preprint
Publié: 2024
Sujets:
Accès en ligne:
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866916125829234688
author Aboagye, Prince
Zheng, Yan
Wang, Junpeng
Saini, Uday Singh
Dai, Xin
Yeh, Michael
Fan, Yujie
Zhuang, Zhongfang
Jain, Shubham
Wang, Liang
Zhang, Wei
author_facet Aboagye, Prince
Zheng, Yan
Wang, Junpeng
Saini, Uday Singh
Dai, Xin
Yeh, Michael
Fan, Yujie
Zhuang, Zhongfang
Jain, Shubham
Wang, Liang
Zhang, Wei
contents The emergence of pre-trained models has significantly impacted Natural Language Processing (NLP) and Computer Vision to relational datasets. Traditionally, these models are assessed through fine-tuned downstream tasks. However, this raises the question of how to evaluate these models more efficiently and more effectively. In this study, we explore a novel approach where we leverage the meta-features associated with each entity as a source of worldly knowledge and employ entity representations from the models. We propose using the consistency between these representations and the meta-features as a metric for evaluating pre-trained models. Our method's effectiveness is demonstrated across various domains, including models with relational datasets, large language models and image models.
format Preprint
id arxiv_https___arxiv_org_abs_2401_02987
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Has Your Pretrained Model Improved? A Multi-head Posterior Based Approach
Aboagye, Prince
Zheng, Yan
Wang, Junpeng
Saini, Uday Singh
Dai, Xin
Yeh, Michael
Fan, Yujie
Zhuang, Zhongfang
Jain, Shubham
Wang, Liang
Zhang, Wei
Computation and Language
Artificial Intelligence
The emergence of pre-trained models has significantly impacted Natural Language Processing (NLP) and Computer Vision to relational datasets. Traditionally, these models are assessed through fine-tuned downstream tasks. However, this raises the question of how to evaluate these models more efficiently and more effectively. In this study, we explore a novel approach where we leverage the meta-features associated with each entity as a source of worldly knowledge and employ entity representations from the models. We propose using the consistency between these representations and the meta-features as a metric for evaluating pre-trained models. Our method's effectiveness is demonstrated across various domains, including models with relational datasets, large language models and image models.
title Has Your Pretrained Model Improved? A Multi-head Posterior Based Approach
topic Computation and Language
Artificial Intelligence
url https://arxiv.org/abs/2401.02987