A Unified Model of Compositional and Distributional Semantics: Theory and Applications

Lead Research Organisation: University of York
Department Name: Computer Science

Abstract

Abstracts are not currently available in GtR for all funded research. This is normally because the abstract was not required at the time of proposal submission, but may be because it included sensitive information such as personal details.
 
Description The mathematics behind models of language, logic, quantum physics, and computation have a common core. Transferring tools from one field to another allows us to extend models of meaning beyond simple noun phrases, to include words with logical or structural meanings. The application of tools from theoretical and quantum computing to models of language, developed within the project, has resulted in drastic simplification of the complexity of using these models. These simplifications also have useful applications in other fields of computer science and mathematics.
Exploitation Route The findings provide the mathematical and logical machinery for researchers and programmers seeking more comprehensive tools to analyze natural language, and provide a route to making such tools significantly more efficient and less computationally costly. They also demonstrate that other proposed approaches cannot be fruitful for fundamental structural reasons.
Sectors Digital/Communication/Information Technologies (including Software)
URL http://arxiv.org/abs/1303.3170
 
Description Knowledge Transfer Partnership
Amount £470,000 (GBP)
Funding ID KTP010852 
Organisation Innovate UK 
Sector Public
Country United Kingdom
Start 05/2018 
End 04/2021
 
Description FET Open EU Grant Proposal 
Organisation National University of Distance Education
Country Spain 
Sector Academic/University 
PI Contribution We have intiated joint research with UNED to continue the work undertaken in the current project. We have submitted an FET Open grant proposal which has been possible directly as a result of the current project.
Collaborator Contribution They have contributed to the writing of the FET Open grant proposal
Impact None yet
Start Year 2014
 
Description FET Open EU Grant Proposal 
Organisation University of the Basque Country
Country Spain 
Sector Academic/University 
PI Contribution We have intiated joint research with UNED to continue the work undertaken in the current project. We have submitted an FET Open grant proposal which has been possible directly as a result of the current project.
Collaborator Contribution They have contributed to the writing of the FET Open grant proposal
Impact None yet
Start Year 2014
 
Title Dependency based embeddings 
Description We provide dependency based word embeddings and related code for a skipgram variant word embedding that utilizes additional information from dependency graphs. This can be employed in a broad range of natural language processing applications such sentiment analysis, question answering and information retrieval. 
Type Of Technology Software 
Year Produced 2016 
Open Source License? Yes  
Impact We also provide experimental results that show that dependency based embeddings can outperform standard window based embeddings in many natural language processing tasks. Specifically, for three different classification methods: a Support Vector Machine, a Convolutional Neural Network and a Long Short Term Memory Network; the use of our dependency based embeddings improve on question classification, sentiment analysis and relation classification tasks. 
URL http://www.cs.york.ac.uk/nlp/extvec