检索结果-内蒙古大学图书馆

An improved hierarchical neural network model with local and global feature matching for script event prediction

EXPERT SYSTEMS WITH APPLICATIONS 2025年 259卷

作者： Zhou, Pengpeng Wu, Bin Wang, Caiyong He, Longzhu China Acad Civil Aviat Sci & Technol Beijing 100028 Peoples R China Beijing Univ Posts & Telecommun Sch Comp Sci Beijing 100876 Peoples R China Beijing Univ Civil Engn & Architecture Beijing Key Lab Robot B & Funct Res Beijing 100044 Peoples R China

script event prediction aims to predict subsequent events based on given context event sequences, requiring a deep understanding of script contexts. Many hierarchical neural network models have been proposed for this purpose. However, existing models often neglect high-level context modeling, while primarily focusing on low-level event arguments and middle-level event relations. To address this issue, we propose an improved hierarchical neural network model that integrates local and global feature matching for script event prediction. First, we extract low-level and middle-level event features using an event encoding layer and an event relation layer, respectively. In particular, the event encoding layer employs stacked transformers with learnable positional encoding to comprehensively capture the connections between event arguments. Then instead of previous local score-level event matching, we further introduce a novel feature-level event matching layer to globally match the context event chain and candidate events. This layer matches the candidate event associated with each context event based on multiple dimensionless similarity measures and aggregates these local matching vectors into a global matching vector using an attention mechanism, thereby capturing holistic contextual information. Finally, the global matching vector is fed into an event prediction layer comprising a classification network ( e.g. , a multi-layered perceptron) to compute the relatedness score. By stacking four bottom-up layers, our model learns multi-level event interactions and deeply understands the script context, resulting in more accurate event predictions. Experimental results on the New York Times corpus demonstrate that our model outperforms state-of-the-art baselines, while being computationally efficient, stable, and quick to converge during training.

关键词： script event prediction Transformer Graph neural network Local and global feature matching Multiple-Choice Narrative Cloze

来源：评论

学校读者我要写书评

暂无评论

Commonsense knowledge enhanced event graph representation learning for script event prediction

引用

MACHINE LEARNING 2025年第3期114卷 1-19页

作者： Li, Xiang Jiang, Xinxi Zhou, Qifeng Xiamen Univ Sch Aerosp Engn Xiamen 361106 Fujian Peoples R China

script event prediction plays an important role in many artificial intelligence applications. A key challenge in this task is accurately understanding the correlation between events and then inferring the subsequent events. Benefiting from exploring the rich connections among events, recently some event graph-based methods have achieved remarkable success. However, limited by the abstract representation of events, it is often difficult to derive the event relationships fully and precisely from the original event graphs. In this paper, we propose a novel framework, called Commonsense knowledge-enhanced event evolutionary graph (CEEG), to remedy this problem. CEEG constructs two core modules, event evolutionary graph and event commonsense graph, the former incorporates connections among the predicate events and object events to find the potential information among events, while the latter utilizes external commonsense knowledge to further enhance the understanding of event relationships. Pre-trained model RoBERTa and Graph Neural Network are also integrated into the framework to obtain more accuracy event nodes depiction and rich event graph information. Comprehensive experimental results on the benchmark dataset show the effectiveness of the proposed framework.

关键词： script event prediction event graph Commonsense knowledge Graph representation learning

来源：评论

学校读者我要写书评

暂无评论

script event prediction via Multilingual event Graph Networks

引用

ACM TRANSACTIONS ON ASIAN AND LOW-RESOURCE LANGUAGE INFORMATION PROCESSING 2023年第2期22卷 1-17页

作者： Zhou, Bo Chen, Yubo Liu, Kang Zhao, Jun Univ Chinese Acad Sci Sch Artificial Intelligence 19 A Yuquan Rd Beijing 100049 Peoples R China Chinese Acad Sci Natl Lab Pattern Recognit Inst Automat 95 Zhongguancun East Rd Beijing 100190 Peoples R China Beijing Acad Artificial Intelligence 1 Zhongguancun East Rd Beijing 100084 Peoples R China

Predicting what happens next in text plays a critical role in building NLP applications. Many methods including count-based and neural-network-based have been proposed to tackle the task called script event prediction: predicting the most suitable subsequent event from a candidate list given a chain of narrative events (context). However, two problems including event ambiguity and evidence bias hinder the performance of these monolingual approaches. The former means that some events in the event chain are ambiguous. The latter means that both the wrong and correct candidate events can obtain sufficient support from the event context. In this article, we propose a novel multilingual approach to address two issues simultaneously. Specifically, to alleviate the event ambiguity problem, we project the monolingual event chains to parallel crosslingual event chains, which can provide complementary information for monolingual event disambiguation. To deal with the evidence bias problem, we construct two monolingual event graphs and a cross-lingual event aligned graph to fully explore connections between events. What's more, we design a graph attention mechanism to model the confidence of the complement clues, which controls the information integration from various languages. By modeling the events with graphs instead of pairs or chains, the model can compare the candidate subsequent events simultaneously and choose the more suitable subsequent event as the final answer. Extensive experiments were conducted on the widely used New York Times corpus for script event prediction task and experimental results show that our approach outperforms previous models.

关键词： script event prediction multilingual event graph network

来源：评论

学校读者我要写书评

暂无评论

script event prediction method based on self-attention mechanism and graph representation learning 6

Script event prediction method based on self-attention mecha...

引用

6th IEEE Advanced Information Technology, Electronic and Automation Control Conference (IEEE IAEAC)

作者： Hu, Meng Bai, Lu Yang, Mei Northwestern Polytech Univ Sch Software Xian Peoples R China Northwestern Polytech Univ Sch Comp Sci Xian Peoples R China

ISBN: (数字)9781665458641

ISBN: (纸本)9781665458641

script event prediction usually refers to giving a sequence of contextual events and then selecting the most likely subsequent event from multiple candidate events. Usually, script event prediction methods are based on event pairs or event chain methods to build prediction models, but these methods can not fully capture the complex relationship between context events and candidate events Aiming at the two difficulties of how to fully mine the meaning of events in text data to accurately represent events and how to make full use of the potential information between event nodes in narrative event graph to improve the accuracy of prediction tasks, this paper proposes a script event prediction model Bert-SatGNN, which combines Bert pre-training model, structural self-attention mechanism and narrative event graph. Our model introduces Bert pre-training mechanism into script event prediction for the first time so that it can more accurately represent the event nodes input into the prediction model. And use the multi-head structure self-attention module to learn the structural information of the event nodes in the narrative event graph, so as to capture the potential information of evolutionary events with causal logic, and finally combine the structural information and time series information to predict the final events. In this paper, experiments are carried out on the widely used New York Times data set, and the experimental results show that our models are better than the most advanced methods.

关键词： script event prediction attention mechanism graph representation learning pre-training model GRU

来源：评论

学校读者我要写书评

暂无评论

script event prediction Based on Large Model Reflection Mechanism 2

Script Event Prediction Based on Large Model Reflection Mech...

引用

2nd International Conference on Computer Vision and Intelligent Technology

作者： Li, Zhiqiang Zhang, Ming Ma, Yupeng Wu, Shihong Wuhan Inst Technol WIT Sch Comp Sci & Engn Artificial Intelligence Wuhan Peoples R China YGSOFT Incs Zhuhai Peoples R China

ISBN: (纸本)9798331540050;9798331540043

script event prediction (SEP) involves predicting the most plausible next event given a sequence of historical events. This paper introduces ReGen-SEPM (Reflection and Generation for script event prediction Model), a framework combining LLaMA-based generation and a reflection mechanism to improve prediction quality. The generative module produces initial predictions based on context and prompts, while the reflection module evaluates and refines them for semantic consistency and causal plausibility. A self-supervised learning mechanism further enhances the reflection module's effectiveness. Experiments on the New York Times Corpus show that ReGen-SEPM outperforms state-of-the-art baselines in accuracy, semantic similarity, and causal coherence, with ablation studies highlighting the contributions of reflection and self-supervised learning.

关键词： script event prediction large language models reflection mechanism llama

来源：评论

学校读者我要写书评

暂无评论

What happens next? Combining enhanced multilevel script learning and dual fusion strategies for script event prediction

引用

INTERNATIONAL JOURNAL OF INTELLIGENT SYSTEMS 2022年第11期37卷 10001-10057页

作者： Zhou, Pengpeng Wu, Bin Wang, Caiyong Peng, Hao Yue, Juwei Xiao, Song Beijing Univ Posts & Telecommun Sch Comp Sci Beijing Key Lab Intelligent Telecommun Software & Beijing 100876 Peoples R China China Acad Civil Aviat Sci & Technol Beijing Peoples R China Beijing Univ Civil Engn & Architecture Sch Elect & Informat Engn Beijing Peoples R China Beihang Univ Beijing Adv Innovat Ctr Big Data & Brain Comp Sch Cyber Sci & Technol Beijing Peoples R China Chinese Acad Sci Inst Informat Engn Beijing Peoples R China

script event prediction (SEP), aiming at predicting next event from context event sequences (i.e., scripts), has played an important role in many real-world applications such as government decision-making. While most of the existing research only depend on the top-level event prediction, they ignore the influence of other bottom levels or other relationship modeling manners. In this paper, we focus on the problem of SEP via multilevel script learning where the goal of is to explore a multistage, multiprediction and multilevel information fusion model for SEP. This is challenging in (1) simultaneously modeling of the multilevel event relationship semantic information and (2) effectively designing multilevel information fusion strategies. In this paper, we propose a new script event prediction model based on Enhanced Multilevel script learning and Dual Fusion strategies, named EMDF-Net. Specifically, EMDF-Net designs the multilevel (event/chain/segment level) script learning to model both temporal and casual information as well as the rich structural relevance via neural stacking of self-attention mechanism and graph neural networks. Then it proposes dual fusion strategies to fully integrate different-level information by nonlinear feature composition and weighted score fusion. Finally, a deep supervision strategy is utilized to end-to-end train the whole model and provide a good initialization for information fusion. Experimental results on the popular NYT corpus demonstrate the effectiveness and superiority of EMDF-Net.

关键词： attention mechanism deep supervision multilevel information fusion script event prediction

来源：评论

学校读者我要写书评

暂无评论

Integrating rich event-level and schema-level information for script event prediction

引用

INTERNATIONAL JOURNAL OF COMPUTATIONAL SCIENCE AND ENGINEERING 2024年第6期27卷 691-702页

作者： Qin, Wei Luo, Xiangfeng Wang, Hao Shanghai Univ Sch Comp Engn & Sci Shanghai Peoples R China

A script consists of a series of structured event sequences extracted from the texts. Given historical scripts, script event prediction aims to predict the subsequent event. The critical aspect in script event prediction is how to effectively represent events, which plays an important role in making accurate predictions. Most existing methods describe events through verbs and a few core independent arguments (i.e., subject, object, and indirect object), which lack the capability to deal with complex and sparse event data. In this paper, we propose a hierarchical event prediction (HEP) model, which integrates information from both event-level and schema-level. At the event level, HEP enriches the existing event representation with extra arguments (i.e., time and place) and modifiers, which provides in-depth event information. At the schema level, it induces the sparse events into conceptual schema, which improves the model's generalisation ability to make more reasonable predictions. To more effectively integrate these two layers of information, we propose an event multi-layer encoder, which can effectively integrate information between different layers. Experimental results on New York Times and CNN corpora demonstrate the effectiveness and superiority of HEP.

关键词： script event prediction SEP event-level schema-level contrast learning bimodal cross attention BCA

来源：评论

学校读者我要写书评

暂无评论

Multi-level Connection Enhanced Representation Learning for script event prediction 21

Multi-level Connection Enhanced Representation Learning for ...

引用

30th World Wide Web Conference (WWW)

作者： Wang, Lihong Yue, Juwei Guo, Shu Sheng, Jiawei Mao, Qianren Chen, Zhenyu Zhong, Shenghai Li, Chen Coordinat Ctr China Natl Comp Network Emergency Response Tech Team Beijing Peoples R China Beihang Univ Beijing Adv Innovat Ctr Big Data & Brain Comp Beijing Peoples R China Chinese Acad Sci Inst Informat Engn Beijing Peoples R China

ISBN: (纸本)9781450383127

script event prediction (SEP) aims to choose a correct subsequent event from a candidate list, given a chain of ordered context events. event representation learning has been proposed and successfully applied to this task. Most previous methods learning representations mainly focus on coarse-grained connections at event or chain level, while ignoring more fine-grained connections between events. Here we propose a novel framework which can enhance the representation learning of events by mining their connections at multiple granularity levels, including argument level, event level and chain level. In our method, we first employ a masked self-attention mechanism to model the relations between the components of events (i.e. arguments). Then, a directed graph convolutional network is further utilized to model the temporal or causal relations between events in the chain. Finally, we introduce an attention module to the context event chain, so as to dynamically aggregate context events with respect to the current candidate event. By fusing threefold connections in a unified framework, our approach can learn more accurate argument/event/chain representations, and thus leads to better prediction performance. Comprehensive experiment results on public New York Times corpus demonstrate that our model outperforms other state-of-the-art baselines. Our code is available in https://***/YueAWu/MCer.

关键词： event representation learning Attention mechanism script event prediction Graph convolutional network

来源：评论

学校读者我要写书评

暂无评论

What Comes Next and Why? A Staged Encoder-Decoder Architecture for script event prediction

What Comes Next and Why? A Staged Encoder-Decoder Architectu...

引用

International Joint Conference on Neural Networks (IJCNN)

作者： Han, Shuaihu Yang, Guohua Zhang, Dawei Tao, Jianhua Univ Chinese Acad Sci Sch Artificial Intelligence Beijing Peoples R China Chinese Acad Sci Inst Automat MAIS Beijing Peoples R China Tsinghua Univ Dept Automat Beijing Peoples R China Tsinghua Univ Beijing Natl Res Ctr Informat Sci & Technol Beijing Peoples R China

ISBN: (纸本)9798350359329;9798350359312

A script, which describes the evolutionary path of events, is a structured event sequence. script event prediction aims to predict the next event from a sequence of historical events. Current studies favor modeling macroscale information, including event sequence, event unit, and event argument, while ignoring the smallest unit of event, i.e., argument vocabulary, which we refer to as microscale information. To fuse event information from different scales, we propose a staged encoderdecoder architecture (SEDA) for script event prediction. SEDA aggregates microscale information to enhance the representation of event arguments and event units in the event-enhancement stage, and extracts the context of event sequence to predict the most relevant candidate event in the context-extraction stage. Both stages of SEDA adopt an efficient and scalable encoderdecoder architecture. The experimental results demonstrate that the accuracy of SEDA on the MCNC task surpasses that of the current SOTA baseline. Additionally, we empoly a method based on the Shapley value to calculate the importance of different event units in an event sequence, providing a quantitative analysis of the prediction results.

关键词： script event prediction microscale information information fusion quantitative analysis

来源：评论

学校读者我要写书评

暂无评论

NarGNN: Narrative Graph Neural Networks for New script event prediction Problem 18

NarGNN: Narrative Graph Neural Networks for New Script Event...

引用

18th IEEE Int Symp on Parallel and Distributed Proc with Applicat (ISPA) / 10th IEEE Int Conf on Big Data and Cloud Comp (BDCloud) / IEEE Int Symp on Social Comp and Networking (SocialCom) / IEEE Int Conf on Sustainable Comp and Commun (SustainCom)

作者： Yang, Shuang Wang, Fali Zha, Daren Xue, Cong Tang, Zhihao Univ Chinese Acad Sci Sch Cyber Secur Beijing Peoples R China Chinese Acad Sci Inst Informat Engn Beijing Peoples R China

ISBN: (纸本)9781665414852

scripts encode world knowledge that can help text understanding, and script learning is to automatically learn this knowledge from unstructured text. Previous script event prediction only considers predicting subsequent events given an existing event sequence. However, in real life, we want to predict events at any location rather than just the last event, because we can control subsequent events by preventing or promoting intermediate events. To this end, we introduce a novel approach, named NarGNN, which can integrate the factual knowledge and experiential knowledge to predict the intermediate event. Specifically, we first construct an event evolutionary graph from the newswire corpus. Then we use Fact Encoder Layer to encode existing event facts, including the semantic information of the event itself and the sequence information among events. Third, we use Fusion Layer to fuse the graph information containing experiential knowledge and the embeddings of existing facts obtained from the previous layer. Fourth, Attention Layer is used to choose the most reasonable result. Finally, our proposed model is evaluated on widely used New York Times corpus and the results demonstrate significant improvements compared with state-of-the-art methods. Also, it is worth noting that NarGNN can be naturally extended to address the previous task with better performances than other methods.

关键词： narrative event evolutionary graph script event prediction graph neural network

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：