To hop or not, that is the question: Towards effective multi-hop reasoning over knowledge graphs
-
Published:2021-07-26
Issue:5
Volume:24
Page:1837-1856
-
ISSN:1386-145X
-
Container-title:World Wide Web
-
language:en
-
Short-container-title:World Wide Web
Author:
Liao Jinzhi, Zhao XiangORCID, Tang Jiuyang, Zeng Weixin, Tan Zhen
Abstract
AbstractWith the proliferation of large-scale knowledge graphs (KGs), multi-hop knowledge graph reasoning has been a capstone that enables machines to be able to handle intelligent tasks, especially where some explicit reasoning path is appreciated for decision making. To train a KG reasoner, supervised learning-based methods suffer from false-negative issues, i.e., unseen paths during training are not to be found in prediction; in contrast, reinforcement learning (RL)-based methods do not require labeled paths, and can explore to cover many appropriate reasoning paths. In this connection, efforts have been dedicated to investigating several RL formulations for multi-hop KG reasoning. Particularly, current RL-based methods generate rewards at the very end of the reasoning process, due to which short paths of hops less than a given threshold are likely to be overlooked, and the overall performance is impaired. To address the problem, we propose , a revised RL formulation of multi-hop KG reasoning that is characterized by two novel designs—the stop signal and the worth-trying signal. The stop signal instructs the agent of RL to stay at the entity after finding the answer, preventing from hopping further even if the threshold is not reached; meanwhile, the worth-trying signal encourages the agent to try to learn some partial patterns from the paths that fail to lead to the answer. To validate the design of our model , comprehensive experiments are carried out on three benchmark knowledge graphs, and the results and analysis suggest the superiority of over state-of-the-art methods.
Funder
National Natural Science Foundation of China
Publisher
Springer Science and Business Media LLC
Subject
Computer Networks and Communications,Hardware and Architecture,Software
Reference45 articles.
1. Arras, L., Montavon, G., Müller, K., Samek, W.: Explaining recurrent neural network predictions in sentiment analysis. In: WASSA@EMNLP 2017, September 8 2017, pp 159–168, Copenhagen, Denmark (2017) 2. Auer, S., Bizer, C., Kobilarov, G., Lehmann, J., Cyganiak, R., Ives, Z.G.: Dbpedia: A nucleus for a web of open data. In: ISWC 2007 + ASWC 2007, November 11-15, 2007, pp 722–735, Busan, Korea (2007) 3. Balazevic, I., Allen, C., Hospedales, T.M.: Hypernetwork knowledge graph embeddings. In: ICANN 2019, September 17-19, 2019, pp 553–565, Munich, Germany (2019) 4. Bordes, A., Glorot, X., Weston, J., Bengio, Y.: A semantic matching energy function for learning with multi-relational data - application to word-sense disambiguation. Mach. Learn. 94(2), 233–259 (2014) 5. Bordes, A., Usunier, N., García-Durán, A., Weston, J., Yakhnenko, O.: Translating embeddings for modeling multi-relational data. In: NeurIPS 2013, December 5-8, 2013, pp 2787–2795, Lake Tahoe, Nevada, United States (2013)
Cited by
4 articles.
订阅此论文施引文献
订阅此论文施引文献,注册后可以免费订阅5篇论文的施引文献,订阅后可以查看论文全部施引文献
|
|