Ensembling Finetuned Language Models for Text Classification

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Arango, Sebastian Pineda, Janowski, Maciej, Purucker, Lennart, Zela, Arber, Hutter, Frank, Grabocka, Josif
Format: Preprint
Veröffentlicht: 2024
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866916455782547456
author Arango, Sebastian Pineda
Janowski, Maciej
Purucker, Lennart
Zela, Arber
Hutter, Frank
Grabocka, Josif
author_facet Arango, Sebastian Pineda
Janowski, Maciej
Purucker, Lennart
Zela, Arber
Hutter, Frank
Grabocka, Josif
contents Finetuning is a common practice widespread across different communities to adapt pretrained models to particular tasks. Text classification is one of these tasks for which many pretrained models are available. On the other hand, ensembles of neural networks are typically used to boost performance and provide reliable uncertainty estimates. However, ensembling pretrained models for text classification is not a well-studied avenue. In this paper, we present a metadataset with predictions from five large finetuned models on six datasets, and report results of different ensembling strategies from these predictions. Our results shed light on how ensembling can improve the performance of finetuned text classifiers and incentivize future adoption of ensembles in such tasks.
format Preprint
id arxiv_https___arxiv_org_abs_2410_19889
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Ensembling Finetuned Language Models for Text Classification
Arango, Sebastian Pineda
Janowski, Maciej
Purucker, Lennart
Zela, Arber
Hutter, Frank
Grabocka, Josif
Computation and Language
Machine Learning
Finetuning is a common practice widespread across different communities to adapt pretrained models to particular tasks. Text classification is one of these tasks for which many pretrained models are available. On the other hand, ensembles of neural networks are typically used to boost performance and provide reliable uncertainty estimates. However, ensembling pretrained models for text classification is not a well-studied avenue. In this paper, we present a metadataset with predictions from five large finetuned models on six datasets, and report results of different ensembling strategies from these predictions. Our results shed light on how ensembling can improve the performance of finetuned text classifiers and incentivize future adoption of ensembles in such tasks.
title Ensembling Finetuned Language Models for Text Classification
topic Computation and Language
Machine Learning
url https://arxiv.org/abs/2410.19889