Winograd schemas in portuguese.

• Ano de defesa: 2020
• Autor(a) principal: Melo, Gabriela Souza de
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: eng
• Instituição de defesa: Biblioteca Digitais de Teses e Dissertações da USP
• Palavras-chave em Português: Análise de texto; Inteligência artificial; Linguagem natural; Sintaxe e semântica da linguagem natural
• Link de acesso: https://www.teses.usp.br/teses/disponiveis/3/3141/tde-14012021-124730/

The Winograd Schema Challenge has become a common benchmark for question answering and natural language processing, having been originally developed in the English language. In order to stimulate the development of Natural Language Processing in Portuguese, we have developed a set of Winograd Schemas in Portuguese. We have also adapted solutions proposed for the English-based version of the challenge so as to have an initial model for usage in its Portuguese-based version. To do so, we created a language model for Portuguese based on a set of Wikipedia documents. In order to evaluate the impact of the increase in the language model capacity in the results for the Portuguese challenge, we tested training of the model with the usage of a myriad of techniques that have previously led to state-of-the-art results for English-based tasks.