近年来,随着互联网技术和应用模式的迅猛发展,互联网数据规模爆炸式增长,其中包含大量带有时序信息的动态事件知识。为了建模这类动态事件知识,时序知识图谱在传统知识图谱的基础上引入时间信息,以带时间戳的知识图谱序列刻画这类知识。时序知识图谱推理任务旨在根据过去发生的事件四元组(主语实体,关系(事件类型),宾语实体,时间戳)预测未来发生的事件。为此,模型需要充分建模实体的历史演化过程。然而,巨大的实体数目以及它们对应的大量历史事件给时序知识图谱推理任务带来了巨大挑战。为了降低待建模历史的规模,已有方法选择建模查询实体的长程历史或者全部实体的短程历史,都丢失了一部分历史信息。实际上,由于不同实体对于一个查询的相关程度不同,模型需要更充分地建模相关实体的历史信息。基于此,该文提出了基于多历史序列联合演化建模的两阶段时序推理模型MENet(Multi-sequence Evolution Network)。具体而言,其在第一阶段采用了一种基于启发式规则的候选实体筛选策略,选择最有可能发生事件的候选实体,从而有效地降低了需要建模的实体数目;在第二阶段,其采用了一个多历史序列联合演化模型: 首先通过组合多个实体各自的长程历史信息,得到需要建模的图序列,进而通过考虑该图序列上同时刻发生事件之间的结构依赖、事件发生的时间数值信息以及不同时刻之间的时序依赖,从而更精准地建模实体演化过程。在三个标准数据集上的实验结果表明,上述模型相比于当前最先进的方法模型具有更好的推理性能。
Abstract
Temporal knowledge graphs integrate temporal information into traditional knowledge graphs and describe such dynamic event knowledge by sequences of knowledge graphs with timestamps. The temporal knowledge graph reasoning task aims to predict future events based on the historical event quadruples (subject entity, relation (event type), object entity, timestamp). To characterize the evolution process of the historical events comprehensively, this paper proposes a two-stage model, called MENet (Multi-sequence Evolution Network), based on jointly evolutional modeling of multiple history sequences. Specifically, in the first candidate entity selection stage, a candidate entity selection strategy is designed via heuristic rules, thus effectively reducing the number of entities to be modeled. In the second stage, it combines the long-term historical sequence of multiple entities to form a graph sequence, and models the evolution process of entities by capturing the structural dependency of concurrent events, the time value information of events, and the temporal dependencies across different timestamps. Experimental results on three standard datasets show that the proposed model outperforms state-of-the-art ones.
关键词
时序推理 /
知识图谱
{{custom_keyword}} /
Key words
temporal reasoning /
knowledge graphs
{{custom_keyword}} /
{{custom_sec.title}}
{{custom_sec.title}}
{{custom_sec.content}}
参考文献
[1] ZOU X H. A survey on application of knowledge graph[J]. Journal of Physics, 2020,1487(1): 012-016.
[2] WARD M D, ANDREAS B, JOSH C,et al. Comparing GDELT and ICEWS event data[J]. Analysis 2013,21(1): 267-297.
[3] JIN W J, QU M, JIN X S, et al. Recurrent event network: Autoregressive structure inference over temporal knowledge graphs[C]//Proceedings of the Conference on Empirical Methods in Natural Language Processing, 2020: 6669-6683.
[4] LI Z X, JIN X L, GUAN S P, et al. Search from history and reason for future: Two-stage reasoning on temporal knowledge graphs[C]//Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing, 2021: 4732-4743.
[5] SUN H H, ZHONG J L, MA Y P, et al. Timetraveler: Reinforcement learning for temporal knowledge graph forecasting[J/OL]. arXiv preprint arXiv: 2109.04101, 2021.
[6] LI Z X, JIN X L, LI W, et al. Temporal knowledge graph reasoning based on evolutional representation learning[C]//Proceedings of the 44th International ACM SIGIR Conference on Research and Development in Information Retrieval, 2021: 408-417.
[7] ZHU C H, CHEN M H, FAN C J, et al. Learning from history: Modeling temporal knowledge graphs with sequential copy-generation networks[C]//Proceedings of the AAAI Conference on Artificial Intelligence, 2021,35(5): 4732-4740.
[8] HAN Z, PENG C, MA Y P,et al. Explainable subgraph reasoning for forecasting on temporal knowledge graphs[C]//Proceedings of the International Conference on Learning Representations, 2020.
[9] HAN Z, ZIFENG D, YUNPU M, et al. Learning neural ordinary equations for forecasting future links on temporal knowledge graphs[C]//Proceedings of the Conference on Empirical Methods in Natural Language Processing, 2021: 8352-8364.
[10] JIANG T S, TIANYU L, TAO G, et al. Towards time-aware knowledge graph completion[C]//Proceedings of COLING, the 26th International Conference on Computational Linguistics: Technical Papers, 2016: 1715-1724.
[11] ALBERTO G D, Sebastijan Dumancˇic＇,et al. Learning sequence encoders for temporal knowledge graph completion[J]. arXiv preprint arXiv: 1809.03202 (2018).
[12] LEBLAY J, MELISACHEW W C. Deriving validity time in knowledge graph[C]//Proceedings of the Web Conference, 2018: 1771-1776.
[13] DASGUPTA S S, SWAYAMBHU N R, et al. Hyte: Hyperplane-based temporally aware knowledge graph embedding[C]//Proceedings of the Conference on Empirical Methods in Natural Language Processing, 2018: 2001-2011.
[14] LACROIX T, GUILLAUME O, NICOLAS U. Tensor decompositions for temporal knowledge base completion[J/OL]. ArXiv preprint arXiv: 2004.04926,2020.
[15] VASHISHTH S, SOUMYA S, VIKRAM N, et al. Composition-based multirelational graph convolutional networks[J/OL]. ArXiv preprint arXiv: 1911.03082,2019.
[16] SHANG C, YUN T, JING H, et al. End-to-end structure-aware convolutional networks for knowledge base completion[C]//Proceedings of the AAAI Conference on Artificial Intelligence, 2019,33(1): 3060-3067.
[17] BOSCHEE E, JENNIFER L, et al. ICEWS coded event data[C]//Proceedings of the Harvard Dataverse 12,2015.
[18] LEETARU K, PHILIP A S. Gdelt: Global data on events, location, and tone, 1979-2012[J]. ISA Annual Convention, Citeseer, 2013,2(4): 1-49.
[19] YANG B, WEN TAU Y, XIAODONG H, et al. Embedding entities and relations for learning and inference in knowledge bases[J/OL]. ArXiv preprint arXiv: 1412.6575,2014.
[20] TROUILLON T, JOHANNES W, SEBASTIAN R, et al. Complex embeddings for simple link prediction[C]//Proceedings of the International Conference on Machine Learning, PMLR, 2016: 2071-2080.
[21] DETTMERS T, PASQUALE M I, PONTUS S, et al. Convolutional 2d knowledge graph embeddings[C]//Proceedings of the AAAI Conference on Artificial Intelligence, 2018.
[22] SUN Z Q, DENG Z H, NIE J Y, et al. Rotate: Knowledge graph embedding by relational rotation in complex space[J/OL]. ArXiv preprint arXiv: 1902.10197,2019.
[23] GOEL R, SEYED M K, MARCUS B, et al. Diachronic embedding for temporal knowledge graph completion[C]//Proceedings of the AAAI Conference on Artificial Intelligence, 2020,34(4): 3988-3995.
{{custom_fnGroup.title_cn}}
脚注
{{custom_fn.content}}
基金
国家自然科学基金(62002341, U1911401, 61772501);国防科技创新项目;中国科学院青年创新促进会(20144310)
{{custom_fund}}
PDF(2026 KB)
