Joint Relational Embeddings for Knowledge-based Question Answering

Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing (EMNLP), pages 645–650, October 25-29, 2014, Doha, Qatar. c⃝2014 Association for Computational Linguistics Joint Relational Embeddings for Knowledge-based Question Answering Min-Chul Yang† Nan Duan‡ Ming Zhou‡ Hae-Chang Rim† †Dept. of Computer & Radio Comms. Engineering, Korea University, Seoul, South Korea ‡Microsoft Research Asia, Beijing, China mcyang@nlp.korea.ac.kr {nanduan, mingzhou}@microsoft.com rim@nlp.korea.ac.kr Abstract Transforming a natural language (NL) question into a corresponding logical form (LF) is central to the knowledge-based question answering (KB-QA) task. Un- like most previous methods that achieve this goal based on mappings between lex- icalized phrases and logical predicates, this paper goes one step further and pro- poses a novel embedding-based approach that maps NL-questions into LFs for KB- QA by leveraging semantic associations between lexical representations and KB- properties in the latent space. Experimen- tal results demonstrate that our proposed method outperforms three KB-QA base- line methods on two publicly released QA data sets. 1 Introduction Knowledge-based question answering (KB-QA) involves answering questions posed in natural language (NL) using existing knowledge bases (KBs). As most KBs are structured databases, how to transform the input question into its corre- sponding structured query for KB (KB-query) as a logical form (LF), also known as semantic pars- ing, is the central task for KB-QA systems. Pre- vious works (Mooney, 2007; Liang et al., 2011; Cai and Yates, 2013; Fader et al., 2013; Berant et al., 2013; Bao et al., 2014) usually leveraged map- pings between NL phrases and logical predicates as lexical triggers to perform transformation tasks in semantic parsing, but they had to deal with two limitations: (i) as the meaning of a logical pred- icate often has different natural language expres- sion (NLE) forms, the lexical triggers extracted for a predicate may at times are limited in size; (ii) entities detected by the named entity recognition (NER) component will be used to compose the logical forms together with the logical predicates, so their types should be consistent with the pred- icates as well. However, most NER components used in existing KB-QA systems are independent from the NLE-to-predicate mapping procedure. We present a novel embedding-based KB-QA method that takes all the aforementioned lim- itations into account, and maps NLE-to-entity and NLE-to-predicate simultaneously using sim- ple vector operations for structured query con- struction. First, low-dimensional embeddings of n-grams, entity types, and predicates are jointly learned from an existing knowledge base and from entries <entitysubj, NL relation phrase, entityobj> that are mined from NL texts labeled as KB- properties with weak supervision. Each such en- try corresponds to an NL expression of a triple <entitysubj, predicate, entityobj> in the KB. These embeddings are used to measure the semantic as- sociations between lexical phrases and two prop- erties of the KB, entity type and logical predicate. Next, given an NL-question, all possible struc- tured queries as candidate LFs are generated and then they are ranked by the similarity between the embeddings of observed features (n-grams) in the NL-question and the embeddings …

Content truncated to 500 words