Making Pre-trained Language Models Better Continual Few-Shot Relation Extractors

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Ma, Shengkun, Han, Jiale, Liang, Yi, Cheng, Bo
Natura: Preprint
Pubblicazione: 2024
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866913243991113728
author Ma, Shengkun
Han, Jiale
Liang, Yi
Cheng, Bo
author_facet Ma, Shengkun
Han, Jiale
Liang, Yi
Cheng, Bo
contents Continual Few-shot Relation Extraction (CFRE) is a practical problem that requires the model to continuously learn novel relations while avoiding forgetting old ones with few labeled training data. The primary challenges are catastrophic forgetting and overfitting. This paper harnesses prompt learning to explore the implicit capabilities of pre-trained language models to address the above two challenges, thereby making language models better continual few-shot relation extractors. Specifically, we propose a Contrastive Prompt Learning framework, which designs prompt representation to acquire more generalized knowledge that can be easily adapted to old and new categories, and margin-based contrastive learning to focus more on hard samples, therefore alleviating catastrophic forgetting and overfitting issues. To further remedy overfitting in low-resource scenarios, we introduce an effective memory augmentation strategy that employs well-crafted prompts to guide ChatGPT in generating diverse samples. Extensive experiments demonstrate that our method outperforms state-of-the-art methods by a large margin and significantly mitigates catastrophic forgetting and overfitting in low-resource scenarios.
format Preprint
id arxiv_https___arxiv_org_abs_2402_15713
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Making Pre-trained Language Models Better Continual Few-Shot Relation Extractors
Ma, Shengkun
Han, Jiale
Liang, Yi
Cheng, Bo
Computation and Language
Artificial Intelligence
Continual Few-shot Relation Extraction (CFRE) is a practical problem that requires the model to continuously learn novel relations while avoiding forgetting old ones with few labeled training data. The primary challenges are catastrophic forgetting and overfitting. This paper harnesses prompt learning to explore the implicit capabilities of pre-trained language models to address the above two challenges, thereby making language models better continual few-shot relation extractors. Specifically, we propose a Contrastive Prompt Learning framework, which designs prompt representation to acquire more generalized knowledge that can be easily adapted to old and new categories, and margin-based contrastive learning to focus more on hard samples, therefore alleviating catastrophic forgetting and overfitting issues. To further remedy overfitting in low-resource scenarios, we introduce an effective memory augmentation strategy that employs well-crafted prompts to guide ChatGPT in generating diverse samples. Extensive experiments demonstrate that our method outperforms state-of-the-art methods by a large margin and significantly mitigates catastrophic forgetting and overfitting in low-resource scenarios.
title Making Pre-trained Language Models Better Continual Few-Shot Relation Extractors
topic Computation and Language
Artificial Intelligence
url https://arxiv.org/abs/2402.15713