Colloquium - Understanding and Defending Against Deepfake Threats

April 23, 2021 - 4:00pm to 5:00pm
Zoom - See emails for details
Bimal Viswanath
Virginia Tech | Computer Science

Significant advances have been made in the design and development of deep generative models, a family of machine learning (ML) algorithms capable of learning a given data distribution to produce new variations of that data. A popular example is the Generative Adversarial Network (GAN). While such models have demonstrated impressive utility in non-security applications in healthcare, computer vision, NLP, and other domains, it is important to investigate them through the lens of security. Generative models can produce convincing synthetic content, e.g., photo-realistic images, videos, and text. Such synthetic content or "deepfakes'' can be misused for malicious purposes. Instances of misuse include fake pornographic images and videos, fake images being used to create fake accounts on social media, and fake text articles to spread disinformation. In this talk, I will focus on deepfake threats covering images, videos and text modalities, and discuss the following recent and ongoing work: (1) Our work understanding deepfake threats beyond the web, focusing on attacks in the healthcare domain. (2) Our efforts to understand characteristics of deepfake content in the wild, and robustness of existing defenses. (3) Our ongoing efforts to build robust defenses against deepfake threats.


Bimal Viswanath portrait - submittedBimal Viswanath is an Assistant Professor of Computer Science at Virginia Tech. His research interests are in security, and his ongoing work investigates machine learning systems through the lens of security. He uses data-driven methods to understand new threats raised by advances in machine learning, and also investigates how machine learning can improve security of online services. He obtained his PhD from the Max Planck Institute for Software Systems, and MS from IIT Madras. He also worked as a Researcher at Nokia Bell Labs before starting an academic position.