The role of syntactic dependencies in compositional distributional semantics

This article provides a preliminary semantic framework for Dependency Grammar in which lexical words are semantically defined as contextual distributions (sets of contexts) while syntactic dependencies are compositional operations on word distributions. More precisely, any syntactic dependency uses the contextual distribution of the dependent word to restrict the distribution of the head, and makes use of the contextual distribution of the head to restrict that of the dependent word. The interpretation of composite expressions and sentences, which are analyzed as a tree of binary dependencies, is performed by restricting the contexts of words dependency by dependency in a left-to-right incremental way. Consequently, the meaning of the whole composite expression or sentence is not a single representation, but a list of contextualized senses, namely the restricted distributions of its constituent (lexical) words. We report the results of two large-scale corpus-based experiments on two different natural language processing applications: paraphrasing and compositional translation.

Compositional translation of single-word complex terms using multilingual splitting

Multilingual terminology acquisition from comparable corpora has been attracting the interest of researchers for twenty years, but challenges still remain. Bilingual term alignment, a subtask of multilingual terminology acquisition, requires a pre-processing step, because term structure may differ according to the language. Morphologically constructed terms should be segmented in order to be aligned with their equivalents in other languages. This article addresses the translation of complex terms using a compositional approach. We focus on the pre-processing of such terms and introduce a domain-oriented splitting method that we apply to compound terms belonging to two domains and four languages. The segmentations are used as input to a translation step. We evaluate which percentage of segmentations can be correctly translated by a compositional approach, and which splitting strategy (precision or recall-oriented) performs better. The results are compared to those obtained with the reference segmentations and with a corpus-base splitting method. Our method is close to the reference segmentation and outperforms the corpus-based method.