A Reinforcement Learning-Based Generative Approach for Event Temporal Relation Extraction

Event temporal relation extraction is a crucial task in natural language processing, aimed at recognizing the temporal relations between event triggers in a text. Despite extensive efforts in this area, the existing methods face two main issues. Firstly, the previous models for event temporal relati...

Full description

Saved in:
Bibliographic Details
Main Authors: Zhonghua Wu, Wenzhong Yang, Meng Zhang, Fuyuan Wei, Xinfang Liu
Format: Article
Language:English
Published: MDPI AG 2025-03-01
Series:Entropy
Subjects:
Online Access:https://www.mdpi.com/1099-4300/27/3/284
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849343008580829184
author Zhonghua Wu
Wenzhong Yang
Meng Zhang
Fuyuan Wei
Xinfang Liu
author_facet Zhonghua Wu
Wenzhong Yang
Meng Zhang
Fuyuan Wei
Xinfang Liu
author_sort Zhonghua Wu
collection DOAJ
description Event temporal relation extraction is a crucial task in natural language processing, aimed at recognizing the temporal relations between event triggers in a text. Despite extensive efforts in this area, the existing methods face two main issues. Firstly, the previous models for event temporal relation extraction mainly rely on a classification framework, which fails to output the crucial contextual words necessary for predicting the temporal relations between two event triggers. Secondly, the prior research that formulated natural language processing tasks as text generation problems usually trained the generative models by maximum likelihood estimation. However, this approach encounters potential difficulties when the optimization objective is misaligned with the task performance metrics. To resolve these limitations, we introduce a reinforcement learning-based generative framework for event temporal relation extraction. Specifically, to output the important contextual words from the input sentence for temporal relation identification, we introduce dependency path generation as an auxiliary task to complement event temporal relation extraction. This task is solved alongside temporal relation prediction to enhance model performance. To achieve this, we reformulate the event temporal relation extraction task as a text generation problem, aiming to generate both event temporal relation labels and dependency path words based on the input sentence. To bridge the gap between the optimization objective and task performance metrics, we employ the REINFORCE algorithm to optimize our generative model, designing a novel reward function to simultaneously capture the accuracy of temporal prediction and the quality of generation. Lastly, to mitigate the high variance issue encountered when using the REINFORCE algorithm in multi-task generative model training, we propose a baseline policy gradient algorithm to improve the stability and efficiency of the training process. Experimental results on two widely used datasets, MATRES and TB-DENSE, show that our approach exhibits competitive performance.
format Article
id doaj-art-764459e5bd1b48bc8325d56caf6a1c7d
institution Kabale University
issn 1099-4300
language English
publishDate 2025-03-01
publisher MDPI AG
record_format Article
series Entropy
spelling doaj-art-764459e5bd1b48bc8325d56caf6a1c7d2025-08-20T03:43:11ZengMDPI AGEntropy1099-43002025-03-0127328410.3390/e27030284A Reinforcement Learning-Based Generative Approach for Event Temporal Relation ExtractionZhonghua Wu0Wenzhong Yang1Meng Zhang2Fuyuan Wei3Xinfang Liu4School of Computer Science and Technology, Xinjiang University, Urumqi 830017, ChinaSchool of Computer Science and Technology, Xinjiang University, Urumqi 830017, ChinaSchool of Computer Science and Technology, Xinjiang University, Urumqi 830017, ChinaSchool of Computer Science and Technology, Xinjiang University, Urumqi 830017, ChinaSchool of Computer Science and Technology, Xinjiang University, Urumqi 830017, ChinaEvent temporal relation extraction is a crucial task in natural language processing, aimed at recognizing the temporal relations between event triggers in a text. Despite extensive efforts in this area, the existing methods face two main issues. Firstly, the previous models for event temporal relation extraction mainly rely on a classification framework, which fails to output the crucial contextual words necessary for predicting the temporal relations between two event triggers. Secondly, the prior research that formulated natural language processing tasks as text generation problems usually trained the generative models by maximum likelihood estimation. However, this approach encounters potential difficulties when the optimization objective is misaligned with the task performance metrics. To resolve these limitations, we introduce a reinforcement learning-based generative framework for event temporal relation extraction. Specifically, to output the important contextual words from the input sentence for temporal relation identification, we introduce dependency path generation as an auxiliary task to complement event temporal relation extraction. This task is solved alongside temporal relation prediction to enhance model performance. To achieve this, we reformulate the event temporal relation extraction task as a text generation problem, aiming to generate both event temporal relation labels and dependency path words based on the input sentence. To bridge the gap between the optimization objective and task performance metrics, we employ the REINFORCE algorithm to optimize our generative model, designing a novel reward function to simultaneously capture the accuracy of temporal prediction and the quality of generation. Lastly, to mitigate the high variance issue encountered when using the REINFORCE algorithm in multi-task generative model training, we propose a baseline policy gradient algorithm to improve the stability and efficiency of the training process. Experimental results on two widely used datasets, MATRES and TB-DENSE, show that our approach exhibits competitive performance.https://www.mdpi.com/1099-4300/27/3/284temporal relation extractiongenerative modelsmulti-task learningdependency pathreinforcement learningpolicy gradient method
spellingShingle Zhonghua Wu
Wenzhong Yang
Meng Zhang
Fuyuan Wei
Xinfang Liu
A Reinforcement Learning-Based Generative Approach for Event Temporal Relation Extraction
Entropy
temporal relation extraction
generative models
multi-task learning
dependency path
reinforcement learning
policy gradient method
title A Reinforcement Learning-Based Generative Approach for Event Temporal Relation Extraction
title_full A Reinforcement Learning-Based Generative Approach for Event Temporal Relation Extraction
title_fullStr A Reinforcement Learning-Based Generative Approach for Event Temporal Relation Extraction
title_full_unstemmed A Reinforcement Learning-Based Generative Approach for Event Temporal Relation Extraction
title_short A Reinforcement Learning-Based Generative Approach for Event Temporal Relation Extraction
title_sort reinforcement learning based generative approach for event temporal relation extraction
topic temporal relation extraction
generative models
multi-task learning
dependency path
reinforcement learning
policy gradient method
url https://www.mdpi.com/1099-4300/27/3/284
work_keys_str_mv AT zhonghuawu areinforcementlearningbasedgenerativeapproachforeventtemporalrelationextraction
AT wenzhongyang areinforcementlearningbasedgenerativeapproachforeventtemporalrelationextraction
AT mengzhang areinforcementlearningbasedgenerativeapproachforeventtemporalrelationextraction
AT fuyuanwei areinforcementlearningbasedgenerativeapproachforeventtemporalrelationextraction
AT xinfangliu areinforcementlearningbasedgenerativeapproachforeventtemporalrelationextraction
AT zhonghuawu reinforcementlearningbasedgenerativeapproachforeventtemporalrelationextraction
AT wenzhongyang reinforcementlearningbasedgenerativeapproachforeventtemporalrelationextraction
AT mengzhang reinforcementlearningbasedgenerativeapproachforeventtemporalrelationextraction
AT fuyuanwei reinforcementlearningbasedgenerativeapproachforeventtemporalrelationextraction
AT xinfangliu reinforcementlearningbasedgenerativeapproachforeventtemporalrelationextraction