Identifying the relations that exist between words (or entities) is important for various natural language processing tasks such as, relational search, nounmodifier classification and analogy detection. A popular approach to represent the relations between a pair of words is to extract the patterns in which the words co-occur with from a corpus, and assign each word-pair a vector of pattern frequencies. Despite the simplicity of this approach, it suffers from data sparseness, information scalability and linguistic creativity as the model is unable to handle previously unseen word pairs in a corpus. In contrast, a compositional approach for representing relations between words overcomes these issues by using the attributes of each individual word to indirectly compose a representation for the common relations that hold between the two words. This study aims to compare different operations for creating relation representations from word-level representations. We investigate the performance of the compositional methods by measuring the relational similarities using several benchmark datasets for word analogy. Moreover, we evaluate the different relation representations in a knowledge base completion task.

No abstract

Finding translations for technical terms is an important problem in machine translation. In particular, in highly specialized domains such as biology or medicine, it is difficult to find bilingual experts to annotate sufficient cross-lingual texts in order to train machine translation systems. Moreover, new terms are constantly being generated in the biomedical community, which makes it difficult to keep the translation dictionaries up to date for all language pairs of interest. Given a biomedical term in one language (source language), we propose a method for detecting its translations in a different language (target language). Specifically, we train a binary classifier to determine whether two biomedical terms written in two languages are translations. Training such a classifier is often complicated due to the lack of common features between the source and target languages. We propose several feature space concatenation methods to successfully overcome this problem. Moreover, we study the effectiveness of contextual and character n-gram features for detecting term translations. Experiments conducted using a standard dataset for biomedical term translation show that the proposed method outperforms several competitive baseline methods in terms of mean average precision and top-k translation accuracy.

No abstract

Embedding entities and relations of a knowledge graph in a low-dimensional space has shown impressive performance in predicting missing links between entities. Although progresses have been achieved, existing methods are heuristically motivated and theoretical understanding of such embeddings is comparatively underdeveloped. This paper extends the random walk model (Arora et al., 2016a) of word embeddings to Knowledge Graph Embeddings (KGEs) to derive a scoring function that evaluates the strength of a relation R between two entities h (head) and t (tail). Moreover, we show that marginal loss minimisation, a popular objective used in much prior work in KGE, follows naturally from the loglikelihood ratio maximisation under the probabilities estimated from the KGEs according to our theoretical relationship. We propose a learning objective motivated by the theoretical analysis to learn KGEs from a given knowledge graph. Using the derived objective, accurate KGEs are learnt from FB15K237 and WN18RR benchmark datasets, providing empirical evidence in support of the theory. * Danushka Bollegala holds concurrent appointments as a Professor at University of Liverpool and as an Amazon Scholar. This paper describes work performed at the University of Liverpool and is not associated with Amazon.