Skip to main content

Advertisement

SpringerLink
Log in

Knowledge base completion by learning pairwise-interaction differentiated embeddings

  • Published:
Data Mining and Knowledge Discovery Aims and scope Submit manuscript

Abstract

A knowledge base of triples like (subject entity, predicate relation,object entity) is a very important resource for knowledge management. It is very useful for human-like reasoning, query expansion, question answering (Siri) and other related AI tasks. However, such a knowledge base often suffers from incompleteness due to a large volume of increasing knowledge in the real world and a lack of reasoning capability. In this paper, we propose a Pairwise-interaction Differentiated Embeddings model to embed entities and relations in the knowledge base to low dimensional vector representations and then predict the possible truth of additional facts to extend the knowledge base. In addition, we present a probability-based objective function to improve the model optimization. Finally, we evaluate the model by considering the problem of computing how likely the additional triple is true for the task of knowledge base completion. Experiments on WordNet and Freebase show the excellent performance of our model and algorithm.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Log in via an institution

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Institutional subscriptions

Fig. 1
Fig. 2
Fig. 3

Similar content being viewed by others

Notes

  1. google.com/insidesearch/features/search/knowledge.html, 10-05-2015.

  2. dbpedia.org, 10-05-2015.

  3. geneontology.org, 10-05-2015.

  4. Such as (subject entity, object entity), (subject entity, predicate relation) and (object entity, predicate relation).

  5. For simplicity, we use subject refer to subject entity, predicate refer to predicate relation and object refer to object entity in the next.

  6. Total order is a binary relation (here denoted by \(\ge \)) which is antisymmetric, transitive and total.

  7. \(\forall o_1, o_2 \in E: o_1 \ge _{s,p} o_2 \wedge o_2 \ge _{s,p} o_1 \Rightarrow o_1 = o_2\) (antisymmetry).

  8. \(\forall o_1, o_2, o_3 \in E: o_1 \ge _{s,p} o_2 \wedge o_2 \ge _{s,p} o_3 \Rightarrow o_1 \ge _{s,p} o_3\) (transitivity).

  9. \(\forall o_1,o_2 \in E: o_1 \ne o_2 \Rightarrow o_1 \ge _{s,p} o_2 \vee o_2 \ge _{s,p} o_1\) (totality).

  10. \(f_1,f_2,f_3\) denote the pairwise-interaction functions.

  11. We do not replace both subject entity and object entity with random one at the same time.

  12. \([x]_+\) denotes the positive part of x (i.e. \([x]_+:=max\{0,x\}\)).

  13. The entities of WordNet are denoted by the concatenation of a word, its POS tag and a digital number. The number refers to its sense. E.g. “_payment_NN_1” encodes the first meaning of the noun “payment”.

References

  • Angeli G, Manning CD (2013) Philosophers are mortal: inferring the truth of unseen facts. In: Proceeding of the 2013 Conference on Computational Natural Language Learning, Sofia, Bulgaria, pp 133–142

  • Berant J, Chou A, Frostig R, Liang P (2013) Semantic parsing on Freebase from question-answer pairs. In: Proceeding of the 2013 Conference on Empirical Methods in Natural Language Processing, pp 1533–1544

  • Berant J, Liang P (2014) Semantic parsing via paraphrasing. In: Proceeding of the 2014 Annual Meeting of the Association for Computational Linguistics, pp 1415–1425

  • Bollacker K, Evans C, Paritosh P, Sturge T, Taylor J (2008) Freebase: a collaboratively created graph database for structuring human knowledge. In: Proceeding of the 2008 International Conference on Management of Data, Vancouver, BC, Canada, pp 1247–1250

  • Bordes A, Weston J, Collobert R, Bengio Y (2011) Learning structured embeddings of knowledge bases. In: Proceeding of the 25th Annual Conference on Artificial Intelligence, San Francisco, USA, pp 301–306

  • Bordes A, Glorot X, Weston J, Bengio Y (2012) Joint learning of words and meaning representations for open-text semantic parsing. In: Proceeding of 2012 International Conference on Artificial Intelligence and Statistics, pp 127–135

  • Bordes A, Glorot X, Weston J, Bengio Y (2013a) A semantic matching energy function for learning with multi-relational data. Mach Learn 94(2):233–259

    Article  MathSciNet  Google Scholar 

  • Bordes A, Usunier N, Garcia-Duran A, Weston J, Yakhnenko O (2013b) Translating embeddings for modeling multi-relational data. Proc Adv Neural Inf Process Syst 26:2787–2795

    Google Scholar 

  • Bordes A, Chopra S, Weston J (2014) Question answering with subgraph embeddings. In: Proceeding of the 2014 Conference on Empirical Methods in Natural Language Processing, Doha, Qatar, pp 615–620

  • Castells P, Fernandez M, Vallet D (2007) An adaptation of the vector-space model for ontology-based information retrieval. IEEE Trans Knowl Data Eng 19(2):261–272

    Article  Google Scholar 

  • Fader A, Soderland S, Etzioni O (2011) Identifying relations for open information extraction. In: Proceeding of the 2011 Conference on Empirical Methods in Natural Language Processing, pp 1535–1545

  • Fader A, Zettlemoyer L, Etzioni O (2014) Open question answering over curated and extracted knowledge bases. In: Proceeding of the 2014 International Conference on Knowledge Discovery and Data Mining, pp 1156–1165

  • Graupmann J, Schenkel R, Weikum G (2005) The SphereSearch engine for unified ranked retrieval of heterogeneous XML and web documents. In: Proceeding of the 2005 International Conference on Very Large Data Bases, pp 529–540

  • Huang EH, Socher R, Manning CD, Ng AY (2012) Improving word representations via global context and multiple word prototypes. In: Proceeding of the 2012 Annual Meeting of the Association for Computational Linguistics, pp 873–882

  • Jenatton R, Roux NL, Bordes A, Obozinski GR (2012) A latent factor model for highly multi-relational data. Proc Adv Neural Inf Process Syst 25:3167–3175

    Google Scholar 

  • Miller GA (1995) WordNet: a lexical database for English. Commun ACM 38(11):39–41

    Article  Google Scholar 

  • Mikolov T, Sutskever I, Chen K, Corrado G, Dean J (2013) Distributed representations of words and phrases and their compositionality. In: Proceeding of Advances in Neural Information Processing Systems 26:3111–3119

  • Ng V, Cardie C (2002) Improving machine learning approaches to coreference resolution. In: Proceeding of the 2002 Annual Meeting of the Association for Computational Linguistics, pp 104–111

  • Rendle S, Marinho LB, Nanopoulos A, Schmidt-Thieme L (2009) Learning optimal ranking with tensor factorization for tag recommendation. In: Proceeding of the 2009 International Conference on Knowledge Discovery and Data Mining, pp 727–736

  • Robbins H, Monro S (1951) A stochastic approximation method. Ann Math Stat 22:400–407

    Article  MathSciNet  MATH  Google Scholar 

  • Snow R, Jurafsky D, Ng AY (2005) Learning syntactic patterns for automatic hypernym discovery. In: Proceeding of Advances in Neural Information Processing Systems 17, MIT Press, Cambridge, MA, pp 1297–1304

  • Socher R, Chen D, Manning CD, Ng AY (2013) Reasoning with neural tensor networks for knowledge base completion. Proc Adv Neural Inf Process Syst 26:926–934

    Google Scholar 

  • Suchanek FM, Kasneci G, Weikum G (2007) Yago: a core of semantic knowledge. In: Proceeding of the 2007 International Conference on World Wide Web, pp 697–706

  • Sutskever I, Salakhutdinov R, Tenenbaum J (2009) Modelling relational data using bayesian clustered tensor factorization. In: Proceeding of Advances in Neural Information Processing Systems 22:1821–1828

  • Vallet D, Fernandez M, Castells P (2005) An ontology-based information retrieval model. In: The Semantic Web: Research and Applications. Springer, Berlin Heidelberg, pp 455–470

  • Wang Z, Zhang J, Feng J, Chen Z (2014) Knowledge graph embedding by translating on hyperplanes. In: Proceedings of the 28th AAAI Conference on Artificial Intelligence, pp 1112–1119

  • Weston J, Bordes A, Yakhnenko O, Usunier N (2013) Connecting language and knowledge bases with embedding models for relation extraction. In: Proceeding of 2013 Conference on Empirical Methods in Natural Language Processing, pp 1366–1371

  • Yao X, Durme BV (2014) Information extraction over structured data: Question answering with freebase. In: Proceeding of the 2014 Annual Meeting of the Association for Computational Linguistics, Baltimore, Maryland, USA, pp 956–966

Download references

Acknowledgments

This work was supported by the Natural Science Foundation of China under Grant No. 61300080, No. 61273217, the 111 Project under Grant No. B08004 and FP7 MobileCloud Project under Grant No. 612212. The authors are partially supported by the Key project of China Ministry of Education under Grant No. MCM20130310, Huawei’s Innovation Research Program and Postgraduate Innovation Fund of SICE, BUPT, 2015. We are thankful to the anonymous reviewers of DMKD whose comments helped us improving this work.

Author information

Authors and Affiliations

  1. Beijing University of Posts and Telecommunications, Beijing, China

    Yu Zhao, Sheng Gao & Jun Guo

  2. LIP6, Universit Pierre et Marie Curie, Paris, France

    Patrick Gallinari

Authors
  1. Yu Zhao
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Sheng Gao
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Patrick Gallinari
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. Jun Guo
    View author publications

    You can also search for this author in PubMed Google Scholar

Corresponding authors

Correspondence to Yu Zhao, Sheng Gao or Jun Guo.

Additional information

Responsible editors: Joao Gama, Indre Zliobaite, Alipio Jorge, Concha Bielza.

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Zhao, Y., Gao, S., Gallinari, P. et al. Knowledge base completion by learning pairwise-interaction differentiated embeddings. Data Min Knowl Disc 29, 1486–1504 (2015). https://doi.org/10.1007/s10618-015-0430-1

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s10618-015-0430-1

Keywords

Search

Navigation