Compositional matrix-space models of language: Definitions, properties, and learning methods. (9th January 2023)

• Record Type: Journal Article
• Authors: Asaadi, Shima; Giesbrecht, Eugenie; Rudolph, Sebastian
• Is Part Of: Natural language engineering. Volume 29:Number 1(2023)
• Page Start: 32
• Page End: 80
• Publication Date: 2023-01-09
• Subjects: Compositionality -- Matrix-space model -- Sentiment analysis -- Word representation learning -- Compositionality prediction; Natural language processing (Computer science) -- Periodicals; Software engineering -- Periodicals; 006.35
• DOI: 10.1017/S1351324921000206
• Languages: English
• ISSNs: 1351-3249

Abstract

Abstract: We give an in-depth account of compositional matrix-space models (CMSMs), a type of generic models for natural language, wherein compositionality is realized via matrix multiplication. We argue for the structural plausibility of this model and show that it is able to cover and combine various common compositional natural language processing approaches. Then, we consider efficient task-specific learning methods for training CMSMs and evaluate their performance in compositionality prediction and sentiment analysis.