Enriching Portuguese Word Embeddings with Visual Information

• Bernardo Scapini Consoli PUCRS https://orcid.org/0000-0003-0656-511X
• Renata Vieira University of Évora https://orcid.org/0000-0003-2449-5477

Resumo

This work focuses on the enrichment of existing Portuguese word embeddings with visual information in the form of visual embeddings. This information was extracted from images portraying given vocabulary terms and imagined visual embeddings learned for terms with no image data. These enriched embeddings were tested against their text-only counterparts in common NLP tasks. The results show an increase in performance for several tasks, which indicates that visual information fusion for word embeddings can be useful for word embedding based NLP tasks.

Palavras-chave: Word embedding models, Multimodality, Portuguese language