people
Research group members
Group Leader
MSU Institute for Artificial Intelligence
ISP RAS Research Center for Trusted AI
Dr. Anastasia Antsiferova is a Senior Research Scientist and Group Leader specializing in trusted AI, video quality metrics, and adversarial robustness. She was recognized as a finalist of “Leaders of AI 2024” in Russia. She has developed and supervised a research group in robust video quality metrics, certification methods, and empirical defenses for neural networks.
She led a team of 24 students conducting cutting-edge research in image/video quality measurement and adversarial robustness of neural networks. Her work includes developing an attack on the VMAF quality metric, which was later integrated into Google’s libaom (AV1) encoder, prompting Netflix to launch a new, more robust metric, VMAF NEG. She also founded two startups, NeuroTechSoft and Effective Video Transcoding.
Dr. Antsiferova has published over 20 peer-reviewed research papers in top conferences, including NeurIPS, ICML, ICLR, and AAAI.