Experimental Support for a Categorical Compositional Distributional Model of Meaning
arXiv:1106.4058
Abstract
Modelling compositional meaning for sentences using empirical distributional methods has been a challenge for computational linguists. We implement the abstract categorical model of Coecke et al. (arXiv:1003.4394v1 [cs.CL]) using data from the BNC and evaluate it. The implementation is based on unsupervised learning of matrices for relational words and applying them to the vectors of their arguments. The evaluation is based on the word disambiguation task developed by Mitchell and Lapata (2008) for intransitive sentences, and on a similar new experiment designed for transitive sentences. Our model matches the results of its competitors in the first experiment, and betters them in the second. The general improvement in results with increase in syntactic complexity showcases the compositional power of our model.
11 pages, to be presented at EMNLP 2011, to be published in Proceedings of the 2011 Conference on Empirical Methods in Natural Language Processing
Cited by in corpus (41)
- QNLP in Practice: Running Compositional Models of Meaning on a Quantum Computer
- Modelling, Visualising and Summarising Documents with a Single Convolutional Neural Network
- The Frobenius anatomy of word meanings I: subject and object relative pronouns
- Representation Learning for Natural Language Processing
- Tensor Regression Networks
- Lambek vs. Lambek: Functorial Vector Space Semantics and String Diagrams for Lambek Calculus
- Quantum Algorithms for Compositional Natural Language Processing
- A Study of Entanglement in a Categorical Framework of Natural Language
- Towards a Formal Distributional Semantics: Simulating Logical Calculi with Tensors
- Experimenting with Transitive Verbs in a DisCoCat
- The Frobenius anatomy of word meanings II: possessive relative pronouns
- Category-Theoretic Quantitative Compositional Distributional Models of Natural Language Semantics
- Quantum Mathematics in Artificial Intelligence
- Foundations for Near-Term Quantum Natural Language Processing
- "Not not bad" is not "bad": A distributional account of negation
- A quantum teleportation inspired algorithm produces sentence meaning from word meaning and grammatical structure
- A CCG-Based Version of the DisCoCat Framework
- Evaluating Neural Word Representations in Tensor-Based Compositional Settings
- Using Sentence Plausibility to Learn the Semantics of Transitive Verbs
- Distributional Inclusion Hypothesis for Tensor-based Composition
- Meaning updating of density matrices
- The Mechanism of Additive Composition
- Investigating the Role of Prior Disambiguation in Deep-learning Compositional Models of Meaning
- Modeling Semantic Compositionality with Sememe Knowledge
- Dual Density Operators and Natural Language Meaning
- Uniqueness of Composition in Quantum Theory and Linguistics
- Translating and Evolving: Towards a Model of Language Change in DisCoCat
- Talking Space: inference from spatial linguistic meanings
- Incremental Monoidal Grammars
- Conversational Negation using Worldly Context in Compositional Distributional Semantics
- Contextualized word senses: from attention to compositionality
- CBOW Is Not All You Need: Combining CBOW with the Compositional Matrix Space Model
- Applying Distributional Compositional Categorical Models of Meaning to Language Translation
- Categorical Vector Space Semantics for Lambek Calculus with a Relevant Modality (Extended Abstract)
- Functorial Language Games for Question Answering
- A Categorical Compositional Distributional Modelling for the Language of Life
- Grounded learning for compositional vector semantics
- Pregroup Grammars, their Syntax and Semantics
- Complexity of Grammar Induction for Quantum Types
- Density Matrices with Metric for Derivational Ambiguity
- Categorical Vector Space Semantics for Lambek Calculus with a Relevant Modality