Ethical AI (Machine Unlearning)
Project Overview
This project enables models to forget selected patient records while tracking accuracy changes under compliant data governance.
About This Project
Healthcare machine unlearning workflow where selected patient records can be forgotten while monitoring post-unlearning behavior.
Tech Stack
Team
Mentors
- Basitha
- Heemakshi
Mentees
- Harsha
- Pranav
- Varun
- Vinayak
Methodology
The project follows a structured implementation approach that includes Train baseline prediction model, Execute record-level unlearning/retraining, Visualize pre and post accuracy and predictions, and Apply GDPR/HIPAA aligned anonymization and handling. These steps are executed iteratively to validate assumptions, improve performance, and ensure reliable delivery of the final solution.
Expected Outcome
By the end of this project, the team is expected to deliver Privacy-compliant model lifecycle strategy, and Transparent impact analysis via dashboard. Together, these outcomes reflect both technical feasibility and practical value for demos, evaluation, and future scaling.