Gilda: communication assistant for the Portuguese language

• Grant number: 22/14054-4
• Support Opportunities: Research Grants - Innovative Research in Small Business - PIPE
• Start date: February 01, 2023
• End date: January 31, 2024
• Field of knowledge: Physical Sciences and Mathematics - Computer Science - Computer Systems
• Agreement: FINEP - Centelha SP Program
• Principal Investigator: Iury Cleveston
• Grantee: Iury Cleveston
• CNAE: Desenvolvimento de programas de computador sob encomenda ; Desenvolvimento e licenciamento de programas de computador customizáveis ; Desenvolvimento e licenciamento de programas de computador não-customizáveis
• Associated researchers: Alana de Santana Correia
• Associated research grant(s): 22/16103-2 - Gilda: communication assistant for the Portuguese language, AP.PIPE
• Associated scholarship(s): 23/01198-0 - Support the creation of the corpus for the Portuguese language, focus on the collection of texts, BP.TT; 23/00737-5 - Support creating thesaurus, BP.TT; 23/00762-0 - Data generation support for spell checker creation, BP.TT; + associated scholarships 23/00777-7 - Support the creation of the corpus for the Portuguese language, focus on data processing and annotation, BP.TT; 23/01702-0 - Support for QA of machine learning models, BP.TT; 23/02129-2 - Data generation support for creating clarity corrector for Portuguese language, BP.TT; 23/01294-0 - Data generation support for creating clarity corrector for Portuguese Language, BP.TT; 23/00819-1 - Data generation support for creating style and tone corrector, BP.TT; 23/00703-3 - Support for creating a dictionary of words and meanings, BP.TT - associated scholarships

Abstract

Writing correctly is a problem for most Brazilians due to the complexity of the Portuguese language and the precariousness of teaching. Research on primary education reveals that 40% of Brazilian students have difficulties understanding texts, which impacts future opportunities, such as good professional position, higher income, and conscious participation in political issues (e.g., fake news). Writing properly is a task that mainly impacts students in the literacy process, ENEM, undergraduate, and international students. In addition to teachers, journalists, lawyers, and other professionals who work with the Portuguese language. Although traditional text editors provide spelling corrections, effective written communication involves analysis of semantics, cohesion, coherence, regency, and style appropriate to the context of the communication. We, therefore, propose a writing assistant for the Portuguese language, Gilda.ai (www.gilda.ai). We employ advanced artificial intelligence and data science techniques, such as natural language processing (NLP) in big data, to enable the capture of writing deviations from statistical analysis of spelling, semantics, and style. The most common deviations are long sentences, punctuation and accentuation errors, plural and verb agreement errors, excessive passive voice and clichés, and inappropriate conjunctions, adjectives, and adverbs. In this way, Gilda can detect and suggest advanced writing corrections in real-time, similar to a Portuguese language professional. We make this solution available via the internet, regardless of platform, accessible 24 hours a day. (AU)