Prof. YingLi Tian

Abstract: Modern deep learning architectures, such as Transformers, hybrid CNN–Transformer models, 
and iterative diffusion models, have achieved remarkable success across a wide range of visual 
intelligence tasks. However, these advances often come at the cost of massive parameters, high memory 
requirements, and substantial computational overhead, creating significant barriers to real-world 
deployment and resource-constrained academic research. In this talk, I will present our recent efforts to 
address the fundamental accuracy–efficiency trade-off through two paradigms: 1) Knowledge-Centric 
Efficiency: We develop knowledge distillation frameworks that transfer rich structural representations 
from powerful teacher models to lightweight student networks. Our work spans spectral decoupling for 
spatiotemporal forecasting, cross-image relational knowledge transfer for semantic segmentation, and 
geometric structure distillation for 3D point-cloud understanding. 2) Resource-Centric Efficiency: We 
directly reduce inference costs through model compression techniques, including structural pruning to 
eliminate redundant prediction heads in multimodal tracking systems and adaptive mixed-precision 
quantization strategies that mitigate outlier effects while preserving temporal consistency in low-bit 
diffusion models. Across diverse applications, these approaches substantially reduce latency, memory 
consumption, and computational demands while maintaining near-lossless performance relative to their 
full-scale counterparts, enabling the development of efficient, scalable, and accessible AI systems for 
advanced visual intelligence applications. 
Dr. YingLi Tian, IEEE F

Biography:

Dr. YingLi Tian, IEEE Fellow, is a Distinguished Professor in the 
Department of Electrical Engineering at the City College of New York 
(CCNY) and in the Department of Computer Science at the Graduate 
Center of the City University of New York (CUNY). She is a Fellow of 
the Institute of Electrical and Electronics Engineers (IEEE), the 
International Association of Pattern Recognition (IAPR), and the 
American Association for the Advancement of Science (AAAS), 
recognized for her pioneering contributions to automatic facial expression 
analysis, human activity understanding, and assistive technology. Her 
research on facial expression analysis and database development has had a 
lasting impact on the field and was honored with the Test of Time Award at the IEEE International 
Conference on Automatic Face and Gesture Recognition in 2019. Before joining CCNY in 2008, Dr. Tian 
was a Research Staff Member at the IBM T. J. Watson Research Center, where she led the video analytics 
team. Her current research interests include computer vision, machine learning, artificial intelligence, 
assistive technologies, medical image analysis, and remote sensing. Dr. Tian has authored more than 280 
peer-reviewed journal and conference publications and holds 29 issued U.S. patents.

Email:

Address:New York City