Natural Language Processing
This line aims to develop efficient models for cognitive architectures that are capable of representing, analyzing and understanding human language (complex, ambiguous and diverse), present in texts and speech in Portuguese. Since the main approaches in Natural Language Processing are focused on the English language, one of the objectives of this line is the construction of databases focused on the Portuguese language. For the development of the models, Machine Learning strategies will be investigated to find patterns and understand natural language. Deep neural networks with attention mechanisms are the state of the art for several tasks in Natural Language Processing and will be explored in this line.