面向脚本事件预测的稠密事件图嵌入  

Embedding Dense Event Graph for Script Event Prediction

作　　者：宁佐廷 贾明颐 安莹[3] 段俊文 NING Zuoting;JIA Mingyi;AN Ying;DUAN Junwen(Key Laboratory of Network Crime Investigation of Hunan Provincial Colleges,Hunan Police Academy,Changsha 410138,China;School of Computer Science and Engineering,Central South University,Changsha 410083,China;Big Data Institute,Central South University,Changsha 410083,China)

机构地区：[1]湖南警察学院网络侦查技术湖南省重点实验室,湖南长沙410138 [2]中南大学计算机学院,湖南长沙410083 [3]中南大学大数据研究院,湖南长沙410083

出　　处：《湖南大学学报（自然科学版）》2023年第8期213-222,共10页Journal of Hunan University:Natural Sciences

基　　金：网络犯罪侦查湖南省普通高校重点实验室开放基金资助项目(2020WLFZZC004);湖南省自然科学基金资助项目(2021JJ40783)。

摘　　要：脚本事件预测是指在给定现有上下文事件链的情况下预测后续事件.在现实世界中,不同事件的关系可以自然地表示为图结构,以事件为节点,以时间或因果关系为边.由于语料库规模有限和信息提取工具的能力不足,先前工作中自动构建的事件图会存在稀疏性问题,并且无法集成来自高阶节点的信息以支持多步推理.为了解决这个问题,本文提出使用可学习的多维加权邻接矩阵的稠密事件图(DEG)来解决之前事件图存在的稀疏性问题并表征事件之间的关系强度.为了实现DEG的嵌入表示,本文同时提出了一个通用框架,该框架能够将高阶事件演化信息组合到事件表示中.在多选叙事完形填空(multiple choice narrative cloze,MCNC)和连贯多选叙事完形填空(coherent multiple choice narrative cloze,CMCNC)数据集上进行了实验,结果证明了此框架的有效性.Script Event Prediction refers to predicting the subsequent event based on a given existing chain of context events.In the real world,the relationship of different events can be naturally represented as a graph structure,where events serve as nodes and their temporal or causal relations are depicted as edges.However,previous approaches that automatically constructed event graphs suffer from sparsity problem due to the limited scale of corpus and the incapability of information extraction tools.Moreover,they fail to integrate information from higher order nodes to support multi-step reasoning.To remedy this,we propose a Dense Event Graph(DEG)approach which use a learnable multi-dimensional weighted adjacency matrix to address the sparsity issue and characterize the relation strengths between events.To embed the DEG,we propose a general framework capable of combining high-order event evolution information into the event representations. Experimental results on the multiple choicenarrative cloze( MCNC) and coherent multiple choice narrative cloze( CMCNC) demonstrate the effectiveness of ourapproach.

关 键 词：脚本事件预测 稠密事件图 图卷积网络 事件抽取 

分 类 号：TP399[自动化与计算机技术—计算机应用技术]