Talk by Prof. Wael AbdAlmageed - Vrije Universiteit Amsterdam

Talk title: Truth Decay Plausible Deniability: The Liar’s Dividend

Prof. Wael AbdAlmageed

Thursday, May 2, 2024@ 11h

Room NU-4A25 and online

Host: Filip Ilievski

Biography – Dr. Wael AbdAlmageed is a Tenured Full Professor with the Holcombe Department of Electrical and Computer Engineering at Clemson University. Until 2024, he was a Research Associate Professor at the Department of Electrical and Computer Engineering and the Computer Science Department, and a Research Director and Distinguished Principal Scientist with the Information Sciences Institute, all are units of USC Viterbi School of Engineering. He is the Founding Director of the Visual Intelligence and Multimedia Analytics Laboratory (VIMAL). He obtained his B.S. in electrical engineering in 1994 and his M.S. in computer engineering in 1997 from Mansoura University in Egypt. He also obtained a graduate software engineering diploma in 1997 from the Information Technology Institute in Egypt via a scholarship granted to distinguished graduates from Egyptian universities. He earned his Ph.D. with Distinction from the University of New Mexico in 2003 where he was also awarded the Outstanding Graduate Student award. Prior to joining ISI, Dr. AbdAlmageed was a research scientist with the University of Maryland at College Park, where he led several research efforts for various NSF, DARPA and IARPA programs. His research interests include representation learning, debiased and fair artificial intelligence, and application of artificial intelligence to medicine and healthcare. Dr. AbdAlmageed leads several multi-institution research efforts, including DARPA’s MediFor, GARD and LwLL and IARPA’s Janus, Odin and BRIAR. He has over 1000 publications in top computer vision, machine learning and biometrics conferences and journals, including CVPR, NeurIPS, ICCV, ECCV, ACM MM, PAMI, TBIOM and ICB. Dr. AbdAlmageed is the recipient of 2019 and 2022 USC Information Sciences Institute Achievement Award. His research has also been featured in CNN, Forbes, Glamour UK, Fox News, Time For Kids and PCMag.

Programme

Abstract – Misinformation and disinformation have existed ever since information itself existed, whether it was a state-sponsored actor trying to sow political/civil unrest, manipulate elections or undermine democracies, or simply unintentional ignorance and tribalism. However, rapid advances in artificial intelligence (AI), especially generative models such as generative adversarial networks and encoder-decoder architectures, make it very easy to produce multimedia artifacts (images, video and audio) that show people doing and saying what they never did or said. The widespread use of manipulated and generated multimedia as tools of misinformation on social networking platforms has major societal consequences, including child safety, public opinion and democracy and criminal justice. The prevalence of fabricated multimedia necessitates a new wave of forensic computer vision methods for detecting and localizing image and video manipulations. In this talk, I will give a brief overview of current challenges in multimedia forensics; then, I will present two novel multimedia forensics methods developed by my group. First, I will present a state-of-the-art approach to fingerprinting and attributing AI-generated images based on constructing representation manifolds from image embeddings. Second, I will present a state-of-the-art deepfake detection architecture using intra- and inter-modal self-consistency. Finally, I will wrap up with conclusions and thoughts for future directions in multimedia forensics research and ethical artificial intelligence in general.

Talk by Prof. Wael AbdAlmageed 2 May 2024 11:00 - 12:00

Programme

About Talk by Prof. Wael AbdAlmageed

Quick links

Study

Featured

About VU