I am a Research Associate within the artificial intelligence (safe and ethical) programme under supervision of Adrian Weller at The Alan Turing Institute. Meanwhile, I am a visitor at CleverHans Lab where I was a Postdoctoral Fellow under the supervision of Nicolas Papernot at Vector Institute before joining the Turing.
I have received a PhD in Computer Science from the Queen Mary University of London. During my PhD, I did internships at Inria (with Aurélien Bellet and Emmanuel Vincent on the intersection of differential privacy and audio) and The Alan Turing Institue (with Adria Gascon and Matt Kusner on the intersection of multi-party computation and machine learning). I was also a PhD visitor at Imperial College London working with Hamed Haddadi.
I love to explore the intersection of machine learning, privacy, fairness and image/audio processing.
My research interests and recent works include:
- Designing adversarial examples by considering image/audio semantics (see ColorFool, Mystique, EdgeFool, FilterFool and FoolHD)
- Designing machine learning models tailored to privacy enhancement technologies (see QUOTIENT and PrivEdge)
- Differential privacy for images/audios and graphs (see DPspeech, GAP)
- Privacy attacks (see trap weights)
- Data privacy and edge computing (see DarkneTZ and Private-Feature Extraction)