Knowledge base completion via search-based question answering

Abstract

Over the past few years, massive amounts of world knowledge have been accumulated in publicly available knowledge bases, such as Freebase, NELL, and YAGO. Yet despite their seemingly huge size, these knowledge bases are greatly incomplete. For example, over 70% of people included in Freebase have no known place of birth, and 99% have no known ethnicity. In this paper, we propose a way to leverage existing Web-search-based question-answering technology to fill in the gaps in knowledge bases in a targeted way. In particular, for each entity attribute, we learn the best set of queries to ask, such that the answer snippets returned by the search engine are most likely to contain the correct value for that attribute. For example, if we want to find Frank Zappa's mother, we could ask the query who is the mother of Frank Zappa. However, this is likely to return 'The Mothers of Invention', which was the name of his band. Our system learns that it should (in this case) add disambiguating terms, such as Zappa's place of birth, in order to make it more likely that the search results contain snippets mentioning his mother. Our system also learns how many different queries to ask for each attribute, since in some cases, asking too many can hurt accuracy (by introducing false positives). We discuss how to aggregate candidate answers across multiple queries, ultimately returning probabilistic predictions for possible values for each attribute. Finally, we evaluate our system and show that it is able to extract a large number of facts with high confidence.

Improving Neural Knowledge Base Completion with Cross-Lingual Projections

“…They then associate the recognized patterns to particular KB relations and finally search the corpus for other entity pairs mentioned using the same patterns (Snow et al, 2004;Mintz et al, 2009;Aprosio et al, 2013). A slight modification is the approach by (West et al, 2014) where lexicalized KB relations are posed as queries to a search engine and results are parsed to find pairs of entities between which the initially queried relation holds. Complementary to this, open information extraction methods (Etzioni et al, 2011;Faruqui and Kumar, 2015) extract large amounts of facts from text that can then be used for extending KBs (Dutta et al, 2014).…”

Extracting Relations between Non-Standard Entities using Distant Supervision and Imitation Learning

“…Finally, it is worth noting that the recall for some of the relations is quite low because they only infrequently occur in text, especially in the same sentence as the subject of the relation. These issues can be overcome by performing coreference resolution (Augenstein et al, 2014;Koch et al, 2014), by retrieving more Web pages or improving the information retrieval component of the approach (West et al, 2014) and by combining extractors operating on sentences with other extractors for semi-structured content on Web pages (Carlson et al, 2010).…”

“…In addition, they do not always perform well if they are trained on a different type of text or for a different domain (Derczynski et al, 2015). This issue becomes more important as focus is shifting from using curated text collections such as Wikipedia to texts collected from the Web via search queries (Web-based distant supervision) which can provide better coverage (West et al, 2014).…”

Learning Entity and Relation Embeddings with Entity Description for Knowledge Graph Completion

“…KGs store structured multi-relational data in the form of (head entity, relation, tail entity), which are called facts. However, despite massive facts that can be contained, most of KGs built semiautomatically or collaboratively are still far from complete and always suffered from sparseness [4]. The completion of KGs has become a critical work.…”