Nadarasar Bahavan

Nadarasar Bahavan

Computer Scientist

About Me

I am a PhD Candidate at the University of Melbourne. My research sits at the intersection of robust machine learning and computer vision, specifically focusing on how we can make AI systems more reliable.

🔬 Research Interests

  • Trustworthy AI: Open-set recognition, deepfake detection, and model robustness.
  • Graph Neural Networks (GNNs): Representation learning on complex relational data.
  • Video Understanding: Analyzing temporal dynamics and semantic content in video streams.

🎓 Education & Background

  • PhD in Computer Science/AI, University of Melbourne, Australia.
  • BsC (Hon) in Biomedical Engineering, University of Moratuwa, Sri Lanka.
  • Leisure: When I’m not in the lab, you’ll find me weightlifting 🏋️, reading 📚, or catching a movie 🎬.

📝 Selected Publications

2026

  • Sphor: A Representation Learning Perspective on Open-Set Recognition for Identifying Unknown Classes in Deep Learning Models CVPR Findings 2026 [Paper]
  • Beyond Deepfake vs Real: Facial Deepfake Detection in the Open-Set Paradigm CVPR Workshops 2026 [Paper]

2023

  • Towards an Objective Measurement of Presence, Place Illusion, and Plausibility Illusion in Virtual Reality using Electroencephalography Virtual Reality Journal [Paper]