Embedded Files
Title: Monozygotic Twins and Face Recognition – A Challenge Problem for Generative AI
Abstract: Despite great advances in face recognition accuracy, distinguishing between monozygotic twins is still challenging. An obvious approach to try to improve accuracy of distinguishing twins would be to use large training sets of twins’ images. Assembling such datasets of real persons is inherently difficult. So, is generative AI up to creating datasets for this purpose? This talk will explore the requirements for a positive answer to this question.
Bio: Kevin W. Bowyer is the Schubmehl-Prein Family Professor Emeritus at the University of Notre Dame. He is a Fellow of the AAAS, IEEE, and IAPR. Professor Bowyer has served as Editor-In-Chief of both the IEEE Transactions on Pattern Analysis and Machine Intelligence and the IEEE Transactions on Biometrics, Behavior, and Identity Science, and also as General Chair or Program Chair of conferences such as CVPR, WACV, FG and IJCB.
Homepage: https://www3.nd.edu/~kwb/
Title: End-to-end representation learning in generative models
Abstract: There are lots of different explorations of how to best train image generation models, from pixel-space generation, latent diffusion models, to representation learning alignment, representation auto-encoders and the recent JIT. In this talk, I will first discuss our recent works on end-to-end image generation which trains the VAE and generative model together. Then, I will draw comparisons and insights between our work and existing methods in terms of latent space quality, the usefulness of ImageNet training, and T2I pre-training.
Bio: Liang Zheng is an Associate Professor (tenured) and ARC Future Fellow at the School of Computing, Australian National University. He joined ANU in 2018 and held the CS Futures Fellowship and ARC DECRA Fellowship. He is also a researcher (part-time) at Canva. He received his Ph.D from the Department of Electronic Engineering, Tsinghua University in 2015, and B.S. from the School of Life Science, Tsinghua University in 2010. He has broad interest in computer vision especially generative AI.
Homepage: https://zheng-lab-anu.github.io/
Title: TBD
Bio: Efim Boieru is Senior Manager of Machine Learning Engineering at Incode Technologies, where he leads the R&D team behind Deepsight, a passive face liveness and deepfake detection system. Under his leadership, Incode became the first company globally to pass iBeta Presentation Attack Detection (PAD) Level 3 on both iOS and Android, achieving 0% FAR and 0% FRR against advanced 3D mask attacks. With over a decade of experience in applied machine learning and computer vision, Efim has previously held research leadership roles at Huawei and Bosch research centers. He actively collaborates with academia, including ongoing joint research with Purdue University. Efim holds an MSc from Skoltech and a BSc/MSc in Applied Mathematics and Physics from MIPT, and conducted research in inverse mathematical modeling as a visiting graduate student at MIT.
Homepage: https://www.efim.tech/
About Incode Technologies
Incode is the global leader in digital identity and trust, powered by artificial intelligence and machine learning to deliver secure and seamless user experiences. Its fully automated, proprietary technology—combined with integrations with government databases—enables businesses to verify identities efficiently and accurately, prevent fraud, and streamline onboarding processes.
This unique approach allows organizations to stay ahead of emerging threats, including deepfakes and generative AI–driven attacks, while maintaining the highest standards of security, privacy, and compliance.
Incode has raised more than $250 million in funding and is trusted by global enterprises across financial services, government, healthcare, retail, and more. To learn more, visit www.incode.com.
Page updated
Google Sites
Report abuse