Targeted Distillation for Sentiment Analysis

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Zhang, Yice, Xie, Guangyu, Lin, Jingjie, Bao, Jianzhu, Wang, Qianlong, Zeng, Xi, Xu, Ruifeng
Formato: Preprint
Publicado: 2025
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866908623598256128
author Zhang, Yice
Xie, Guangyu
Lin, Jingjie
Bao, Jianzhu
Wang, Qianlong
Zeng, Xi
Xu, Ruifeng
author_facet Zhang, Yice
Xie, Guangyu
Lin, Jingjie
Bao, Jianzhu
Wang, Qianlong
Zeng, Xi
Xu, Ruifeng
contents This paper explores targeted distillation methods for sentiment analysis, aiming to build compact and practical models that preserve strong and generalizable sentiment analysis capabilities. To this end, we conceptually decouple the distillation target into knowledge and alignment and accordingly propose a two-stage distillation framework. Moreover, we introduce SentiBench, a comprehensive and systematic sentiment analysis benchmark that covers a diverse set of tasks across 12 datasets. We evaluate a wide range of models on this benchmark. Experimental results show that our approach substantially enhances the performance of compact models across diverse sentiment analysis tasks, and the resulting models demonstrate strong generalization to unseen tasks, showcasing robust competitiveness against existing small-scale models.
format Preprint
id arxiv_https___arxiv_org_abs_2503_03225
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Targeted Distillation for Sentiment Analysis
Zhang, Yice
Xie, Guangyu
Lin, Jingjie
Bao, Jianzhu
Wang, Qianlong
Zeng, Xi
Xu, Ruifeng
Computation and Language
This paper explores targeted distillation methods for sentiment analysis, aiming to build compact and practical models that preserve strong and generalizable sentiment analysis capabilities. To this end, we conceptually decouple the distillation target into knowledge and alignment and accordingly propose a two-stage distillation framework. Moreover, we introduce SentiBench, a comprehensive and systematic sentiment analysis benchmark that covers a diverse set of tasks across 12 datasets. We evaluate a wide range of models on this benchmark. Experimental results show that our approach substantially enhances the performance of compact models across diverse sentiment analysis tasks, and the resulting models demonstrate strong generalization to unseen tasks, showcasing robust competitiveness against existing small-scale models.
title Targeted Distillation for Sentiment Analysis
topic Computation and Language
url https://arxiv.org/abs/2503.03225