Advanced search
Start date
Betweenand

MMeaning - multimodal distributional semantic models

Grant number: 16/13002-0
Support type:Regular Research Grants
Duration: October 01, 2016 - October 31, 2018
Field of knowledge:Physical Sciences and Mathematics - Computer Science - Computing Methodologies and Techniques
Principal researcher:Helena de Medeiros Caseli
Grantee:Helena de Medeiros Caseli
Home Institution: Centro de Ciências Exatas e de Tecnologia (CCET). Universidade Federal de São Carlos (UFSCAR). São Carlos , SP, Brazil
Assoc. researchers: Eloize Rossi Marques Seno ; Jander Moreira

Abstract

With the increasing availability of information on the web, processing and retrieving textual and visual information are essential activities in automatic knowledge generation. As most of the information available on the web is made up of text written in natural language and images, process them in a "smart" way necessarily involves understanding (interpretation) of the meaning of the information they convey. One of the most used forms for representing the semantic content are the distributional semantic models, which are based on the distributional hypothesis which states that the meaning of a word is given by its occurrence context. Although the main source for semantic knowledge extraction are the corpora, other sources of extra-linguistic information, such as images, should also be taken into account. The combination of multiple sources of information to generate semantic representations is called multimodal distributional semantic representation. In addition to this new research field, there is the recent interest in distributional representation models based on neural networks, also known as models of deep learning. In this context, this project aims to investigate the use of different sources of knowledge, such as parallel/comparable texts and images in the distributional semantic modeling of natural language texts to enrich the information used in Natural Language Processing and Information Retrieval applications. (AU)

Scientific publications
(References retrieved automatically from Web of Science and SciELO through information on FAPESP grants and their corresponding numbers as mentioned in the publications by the authors)
RODRIGUES DA SILVA, JESSICA; CASELI, HELENA DE M. Sense representations for Portuguese: experiments with sense embeddings and deep neural language models. Language Resources and Evaluation, FEB 2021. Web of Science Citations: 0.

Please report errors in scientific publications list by writing to: cdi@fapesp.br.