2018/10/26 by Mehrnoosh Sadrzadeh, Sadrzadeh, Mehrnoosh, Reinhard Muskens +1
Computer Science · #03B65 #Computation and Language (cs.CL) #FOS: Computer and information sciences #I.2.7 #Logic, Reasoning, and Knowledge #Multi-Agent Systems and Negotiation #Natural Language Processing Techniques #Semantic Web and Ontologies
paper · pdf · doi:10.48550/arxiv.1810.11351
openalex publication_date 2018/10/26 · openalex created_date 2022/08/02 · openalex updated_date 2026/07/28
Vector models of language are based on the contextual aspects of language,\nthe distributions of words and how they co-occur in text. Truth conditional\nmodels focus on the logical aspects of language, compositional properties of\nwords and how they compose to form sentences. In the truth conditional\napproach, the denotation of a sentence determines its truth conditions, which\ncan be taken to be a truth value, a set of possible worlds, a context change\npotential, or similar. In the vector models, the degree of co-occurrence of\nwords in context determines how similar the meanings of words are. In this\npaper, we put these two models together and develop a vector semantics for\nlanguage based on the simply typed lambda calculus models of natural language.\nWe provide two types of vector semantics: a static one that uses techniques\nfamiliar from the truth conditional tradition and a dynamic one based on a form\nof dynamic interpretation inspired by Heim's context change potentials. We show\nhow the dynamic model can be applied to entailment between a corpus and a\nsentence and we provide examples.\n