Active Project
Development of a High-Assurance Platform for Sharing and Trading AI Data and Models
Building a trustworthy platform that lets organizations share and trade AI training data and models without exposing the underlying assets. The work covers provenance and integrity verification, privacy-preserving quality and value assessment, usage control over distributed models, and the governance mechanisms needed for a reliable data and model marketplace.
2026 - Present
·
Data Intelligence & Privacy Lab
·
Ongoing
Active Project
Development of a Hyper-Personalized Media Recommendation AI Agent Based on On-Device Federated Learning with Privacy-Enhanced Context Matrix Computation
Designing a recommendation agent that keeps personal viewing behavior on the user's device. Federated learning trains a shared media recommendation model without collecting raw logs, while privacy-enhanced context matrix computation combines user, content, and situational signals in protected form, enabling hyper-personalized recommendations under strong privacy guarantees.
2026 - Present
·
Data Intelligence & Privacy Lab
·
Ongoing
Active Project
Privacy-Preserving Data Mining
Developing privacy-preserving methods for analyzing sensitive behavioral and productivity data while maintaining practical data utility.
2026 - Present
·
Data Intelligence & Privacy Lab
·
Ongoing
Project
Trustworthy AI for Human-Centered Systems
Exploring trustworthy machine learning and artificial intelligence techniques for human-centered intelligent systems.
2025 - Present
·
Data Intelligence & Privacy Lab
·
Ongoing
Completed Project
Behavioral Data Mining
Investigating data mining methods for discovering meaningful patterns from real-world behavioral and productivity data.
2024 - 2025
·
Data Intelligence & Privacy Lab
·
Completed