[1]车飞虎,张大伟,邵朋朋,等.基于四元数门控图神经网络的脚本事件预测[J].智能系统学报,2023,18(1):138-143.[doi:10.11992/tis.202203042]
CHE Feihu,ZHANG Dawei,SHAO Pengpeng,et al.Script event prediction based on a quaternion-gated graph neural network[J].CAAI Transactions on Intelligent Systems,2023,18(1):138-143.[doi:10.11992/tis.202203042]
点击复制
CHE Feihu,ZHANG Dawei,SHAO Pengpeng,et al.Script event prediction based on a quaternion-gated graph neural network[J].CAAI Transactions on Intelligent Systems,2023,18(1):138-143.[doi:10.11992/tis.202203042]
基于四元数门控图神经网络的脚本事件预测
《智能系统学报》[ISSN 1673-4785/CN 23-1538/TP] 卷:
18
期数:
2023年第1期
页码:
138-143
栏目:
学术论文—自然语言处理与理解
出版日期:
2023-01-05
- Title:
- Script event prediction based on a quaternion-gated graph neural network
- Keywords:
- quaternion; gated graph neural network; event representation; script event prediction; attention mechanism; event evolutionary graph; graph convolution networks; event interaction
- 分类号:
- TP183
- 摘要:
- 脚本事件预测需要考虑两类信息来源：事件间的关联与事件内的交互。针对于事件间的关联，采用门控图神经网络对其进行建模。而对于事件内的交互，采用四元数对事件进行表征，接着通过四元数的哈密顿乘积来捕捉事件4个组成部分之间的交互。提出结合四元数和门控图神经网络来学习事件表示，它既考虑了外部事件图的交互作用，又考虑了事件内部的依赖关系。得到事件表示后，利用注意机制学习上下文事件表示和每个候选上下文表示的相对权值。然后通过权重计算上下文事件表示的和，再计算其与候选事件表示的欧氏距离。最后选择距离最小的候选事件作为正确的候选事件。在纽约时报语库上进行了实验，结果表明，通过多项选择叙事完形填空评价，本文的模型优于现有的基线模型
- Abstract:
- Two types of information sources are essential for script event prediction: the correlation between events and the inner interactions within one event. For the first information source, we use a gated graph neural network to model the correlation between events. For the inner interactions within one event, we use quaternion to model the event, and then we use the Hamilton product of quaternion to capture the inner interactions of four components. We propose to learn event representation by combining quaternion and a gated graph neural network. This approach considers the interaction of external event diagrams and the dependence within an event. After obtaining the event representation, we use an attention mechanism to learn the context event representation and the relative weight of each candidate context representation. Next, we calculate the sum of the context event embeddings through the weights, and then we calculate the Euclidean distance between the context event embedding sum and the candidate event embedding. Finally, we choose the candidate event with the smallest distance as the right candidate event. The results of experiments conducted on the New York Times corpus show that our proposed model is superior to the existing state-of-the-art baseline models through evaluation using a multiple-choice narrative cloze test.
- 参考文献/References:
-
[1] LI ZHONGYANG, DING XIAO, LIU TING. Constructing narrative event evolutionary graph for script event prediction[EB/OL]. (2018?05?14)[2022?03?23].https://arxiv.org/abs/1805.05081.
[2] GEERAERTS, DIKR. Cognitive Linguistics: Basic Readings[M]. Berlin: De Gruyter Mouton, 2008. : 373–400.
[3] RUMELHART D E. Notes on a schema for stories[M]//Representation and Understanding. Amsterdam: Elsevier, 1975: 211?236.
[4] WANG Zhongqing, ZHANG Yue, CHANG C Y. Integrating order information and event relation for script event prediction[C]//Proceedings of the 2017 Conference on Empirical Methods in Natural Language Processing. Copenhagen: Association for Computational Linguistics, 2017: 57–67.
[5] MIZOGUCHI T, YAMADA I. Hyper complex tensor completion with cayley-dickson singular value decomposition[C]//2018 IEEE International Conference on Acoustics, Speech and Signal Processing. Calgary: IEEE, 2018: 3979?3983.
[6] XIANG Min, KANNA S, MANDIC D P. Performance analysis of quaternion-valued adaptive filters in nonstationary environments[J]. IEEE transactions on signal processing, 2018, 66(6): 1566–1579.
[7] ORTOLANI F, COMMINIELLO D, SCARPINITI M, et al. Frequency domain quaternion adaptive filters: Algorithms and convergence performance[J]. Signal processing, 2017, 136: 69–80.
[8] PARCOLLET T, MORCHID M, BOUSQUET P M, et al. Quaternion neural networks for spoken language understanding[C]//2016 IEEE Spoken Language Technology Workshop. San Diego: IEEE, 2016: 362?368.
[9] TITOUAN Parcollet, MIRCO Ravanelli, MOHAMED Morchid, et al. Quaternion recurrent neural networks[EB/OL] (2018?06?12) [2022?03?23].https://arxiv.org/abs/1806.04418.
[10] GAUDET C J, MAIDA A S. Deep quaternion networks[C]//2018 International Joint Conference on Neural Networks. Rio de Janeiro: IEEE, 2018: 1-8.
[11] PEROZZI B, AL-RFOU R, SKIENA S. DeepWalk: online learning of social representations[C]// Proceedings of the 20th ACM SIGKDD international conference on Knowledge discovery and data mining. New York: Association for Computing Machinery, 2014: 701–710.
[12] KIPF T, WELLING M. Semi-supervised classification with graph convolutional networks[EB/OL]. (2016?09?09) [2022?03?23].https://arxiv.org/abs/1609.02907.
[13] WANG Tingwu, ZHOU Yuhao, FIDLER S, et al. Neural graph evolution: towards efficient automatic robot design[EB/OL]. (2016?06?12) [2022?03?23].https://arxiv.org/abs/1906.05370.
[14] LI YUJIA, TARLOW D, BROCKSCHMIDT M, et al. Gated graph sequence neural networks[EB/OL]. (2015?11?17) [2022?03?23].https://arxiv.org/abs/1511.05493.
[15] NATHANAEL Chambers, DAN Jurafsky. Unsupervised learning of narrative event chains[C]//Proceedings of ACL-08: HLT. Columbus: Association for Computational Linguistics, 2008: 789–797.
[16] XU Dongpo, ZHANG Lina, ZHANG Huisheng. Learning algorithms in quaternion neural networks using ghr calculus[J]. Neural network world, 2017, 27(3): 271–282.
[17] CHO K, VAN MERRIENBOER B, BAHDANAU D, et al. On the properties of neural machine translation: encoder-decoder approaches[EB/OL]. (2014?09?03) [2022?03?23].https://arxiv.org/abs/1409.1259.
[18] ZHOU Bo, CHEN Yubo, LIU Kang, et al. Multi-task self-supervised learning for script event prediction[C]//CIKM’21: Proceedings of the 30th ACM International Conference on Information & Knowledge Management. New York: ACM, 2021: 3662?3666.
[19] SCHLICHTKRULL M, KIPF T N, BLOEM P, et al. Modeling relational data with graph convolutional networks[M]//The Semantic Web. Cham: Springer International Publishing, 2018: 593?607.
[20] CHO K, VAN MERRIENBOER B, GULCEHRE C, et al. Learning phrase representations using RNN encoder-decoder for statistical machine translation[EB/OL]. (2014?06?03) [2022?03?23].https://arxiv.org/abs/1406.1078.
[21] BAHDANAU D, CHO K, BENGIO Y. Neural machine translation by jointly learning to align and translate[EB/OL]. (2014?09?01) [2022?03?23].https://arxiv.org/abs/1409.0473.
[22] CURRAN J R, CLARK S, BOS J. Linguistically motivated large-scale NLP with C&C and boxer[C]//ACL’07: Proceedings of the 45th Annual Meeting of the ACL on Interactive Poster and Demonstration Sessions. New York: ACM, 2007: 33?36.
[23] JANS B, BETHARD S, VULI? I, et al. Skip n-grams and ranking functions for predicting script events[C]//EACL’12: Proceedings of the 13th Conference of the European Chapter of the Association for Computational Linguistics. New York: ACM, 2012: 336?344.
[24] MIKOLOV T, SUTSKEVER I, CHEN KAI, et al. Distributed representations of words and phrases and their compositionality[EB/OL]. (2013?10?16)[2022?03?23].https://arxiv.org/abs/1310.4546.
[25] GRANROTH-WILDING M, CLARK S. What happens next? event prediction using a compositional neural network model[C]// Proceedings of the 30th AAAI Conference on Artificial Intelligence and the 28th Innovative Applications of Artificial Intelligence Conference. Phoenix: AAAI Press, 2016, 30(1).
- 相似文献/References:
-
[1]李振娜,王涛,王斌锐,等.基于带约束S型速度曲线的机械手笛卡尔空间轨迹规划[J].智能系统学报,2019,14(4):655.[doi:10.11992/tis.201806025]
LI Zhenna,WANG Tao,WANG Binrui,et al.Trajectory planning for manipulator in Cartesian space based on constrained S-curve velocity[J].CAAI Transactions on Intelligent Systems,2019,14():655.[doi:10.11992/tis.201806025]
备注/Memo
收稿日期:2022-03-23。
基金项目:国家自然科学基金项目(61831022,61901473).
作者简介:车飞虎,博士研究生,主要研究方向为机器学习与数据挖掘;张大伟,副研究员,主要研究方向为模式识别、自然语言处理与知识推理;陶建华,研究员,主要研究方向为语音合成、模式识别、数据挖掘。先后负责和参与国家级项目40余项。主要研究方向为语音合成,模式识别,数据挖掘
通讯作者:陶建华.E-mail:jhtao@nlpr.ia.ac.cn
更新日期/Last Update:
1900-01-01