SDU Responsible and Safe AI Lab
Applied Artificial Intelligence and Data Science Unit
The Mærsk Mc-Kinney Møller Institute · Faculty of Engineering · University of Southern Denmark
SDU RESPONSIBLE AND SAFE AI LAB
Applied Artificial Intelligence and Data Science Unit
The Mærsk Mc-Kinney Møller Institute · Faculty of Engineering · University of Southern Denmark
We build AI people can trust, AI that is private, fair, transparent, safe, and efficient. Our research rests on three pillars, foundational AI, responsible AI and quantum AI, and we bring it into healthcare, the industrial internet of things, life sciences and optics, together with partners in hospitals and industry.
The lab is led by Associate Professor Vinay Chakravarthi Gogineni.
News
| Sep 08, 2026 | Delivered a talk titled “Responsible AI in Clinical and Public Health: Trust, Fairness, and Privacy,” at National Institute of Public Health, Copenhagen. |
|---|---|
| Aug 08, 2026 | The Metric, Not the Model: Anatomy Confounds Attention Faithfulness in Coronary Calcium, has been accepted to MI4MedFM at MICCAI 2026. Congratulations Congratulations Jyothi 👏. |
| Aug 04, 2026 | Our bachelors’ students thesis work “Synthetic Data Augmentation via Stable Diffusion for Polyp Characterisation in Colon Capsule Endoscopy” has been accepted to SASHIMI@MICCAI 2026. Congratulations Florentin Mustafa and Mohamed Wali-Wali 👏. |
| Jun 15, 2026 | Delivered a plenary talk titled “Responsible AI and Mathematical Foundations of Machine Unlearning,” at Scandinavian Conference on AI (SCAI), 2026, Odense. |
| Apr 01, 2026 | Carlsberg Foundation granted 80000 DKK to support Scandinavian Conference on AI (SCAI), 2026! I sincerely thank the Carlsberg Foundation for their generous support. |
Selected Publications
- Responsible AIControlling Path Dependence in Gradient Ascent Unlearning through Forget-Set OrderingIn Forty-Second Annual Conference on Uncertainty in Artificial Intelligence (UAI), 2026
- Responsible AIFINU: Fisher-Informed Noise Injection for Efficient Zero-Shot UnlearningIn 2026 International Joint Conference on Neural Networks (IJCNN), 2026
- Responsible AIRKLU: Redistributive KL Distillation for Efficient Retain-Free Machine UnlearningIn European Symposium on Artificial Neural Networks (ESANN), 2026
- Responsible AIOn the Path Dependence of Gradient Ascent-Based UnlearningIn ICLR Third Workshop on Test-Time Updates (Main Track), 2026
- Responsible AIThe Metric, Not the Model: Anatomy Confounds Attention Faithfulness in Coronary CalciumIn 1st MICCAI Workshop on Mechanistic Interpretability for Medical Foundational Models, 2026
- Applied AISynthetic Data Augmentation via Stable Diffusion for Polyp Characterisation in Colon Capsule EndoscopyIn MICCAI 2026 Workshop on Simulation and Synthesis in Medical Imaging (SASHIMI), 2026
- Responsible AIEfficient Knowledge Deletion from Trained Models Through Layer-wise Partial Machine UnlearningJournal of Machine Learning Research, 2025