Domain Grounded Natural Language Explanations

Project One Liner: Generating natural language explanations in the critical sectors, grounded in domain vocabulary

Status: ongoing

Project Theme: explainability

Project Areas: legal

Team: Gokul S Krishnan, Ambreesh Parthasarathy, Shubham Kashyapi, Krithi Shailya, Chandradithya Jonnalagadda, Bhaarath K, Balaraman Ravindran

Short Description: This project focuses on developing approaches for generating understandable explanations for complex decision-making processes, particularly in high-stakes domains like law and medicine. We make use of graphs to represent the data and the interconnections within data. This is expected to help in unraveling complex connections and patterns for a deeper understanding of multifaceted relationships; Interpretable Applications. Such approaches are particularly beneficial where decisions involve weighing multiple factors, allowing for better analysis and more robust decision-making. The project aims to not only make accurate predictions but also to generate human-understandable explanations for these predictions.

Back to Projects