Research Overview
My primary research endeavors revolve around the critical field of media forensics, with a dedicated focus on addressing the growing challenges posed by audio deepfakes. My work involves developing novel techniques to detect and analyze AI-generated audio by identifying specific, subtle artifacts left behind by synthesis models like neural vocoders. A significant part of this effort is dedicated to creating large-scale, comprehensive forensic datasets that are essential for training and benchmarking robust detection models.
In the contemporary digital landscape, where the widespread proliferation of manipulated media can have profound societal implications, my work aims to bolster the integrity and credibility of digital media. By leveraging advanced algorithms and machine learning, I contribute to projects like the DeepFake-o-meter, an open-source platform for deepfake detection, and investigate multi-modal approaches that combine audio and video analysis for more resilient detection.
Beyond technical development, a cornerstone of my research philosophy is collaboration. I am always eager to connect with experts, researchers, and enthusiasts in media forensics and adjacent domains to foster joint projects and exchange knowledge.
Let's Collaborate
I'm always open to discussing new research ideas and potential collaborations, particularly on topics like segmental speech feature analysis, generative model attribution, and developing robust defenses against adversarial attacks. Feel free to reach out or view my CV for more details.
View My CV