Onat Özdemir
PhD Student · School of Informatics, University of Edinburgh
Informatics Forum
10 Crichton Street
Edinburgh EH8 9AB, UK
I am a PhD student at the School of Informatics, University of Edinburgh, working on multimodal deepfake detection. My work focuses on using vision–language models (VLMs) to detect manipulated and synthetic media in real-world settings.
I am also interested in interpretability and explainable AI (XAI) for vision–language models, especially understanding what these models learn and how they make decisions from both visual and textual information.
Before Edinburgh, I worked as an Applied Scientist at RadiusAI (2022–2025), where I built VLM-based visual search and multi-camera systems for retail environments. I completed my BSc and MSc in Computer Engineering at Middle East Technical University (METU), where my MSc thesis focused on interpretable zero-shot classification using concept bottleneck models.
Research interests. Multimodal deepfake detection · vision–language models · representation learning · interpretability · explainable AI (XAI).
I am always happy to talk about multimodal learning, vision–language models, and interpretability.
news
| Feb 21, 2026 | Our paper Explaining CLIP Zero-shot Predictions Through Concepts has been accepted to CVPR 2026! Project page: oonat.github.io/ezpc. |
|---|---|
| Sep 08, 2025 | Started my PhD at the School of Informatics, University of Edinburgh. |
| Aug 05, 2025 | Completed MSc in Computer Engineering at Middle East Technical University (METU), with a thesis on interpretable zero-shot classification with concept bottleneck models. |
