KENSI: Knowledge-Enhanced NLP for Social Impact

Knowledge-Enhanced Natural Language Machine Learning for Positive Social Impact investigates the development of knowledge-enhanced machine learning methods for natural language processing with a focus on generating positive social impact. By combining structured knowledge sources with modern NLP techniques, the research addresses real-world challenges across misinformation detection, hate speech identification, health communication, and social inequality.

A central motivation of this project is to foster inclusive and responsible AI research, with particular attention to underrepresented languages, communities, and social contexts — including Brazilian Portuguese and the sociotechnical realities of Rio de Janeiro and beyond.

Objectives

• Develop NLP models enriched with domain and commonsense knowledge for socially relevant tasks
• Address challenges in low-resource settings, with emphasis on Brazilian Portuguese
• Apply machine learning to problems such as misinformation detection, harmful content identification, and health text analysis
• Promote diversity in AI research by supporting women researchers in computing in Rio de Janeiro

Funding & Support

This project is financed by FAPERJ (Carlos Chagas Filho Foundation for Research Support of the State of Rio de Janeiro) under the Young Women Researchers in Rio de Janeiro program, covering the period from March 2024 to March 2027.