Yang YOU

Yang YOU is a NUS Presidential Young Professor at the National University of Singapore, Department of Computer Science. He received his PhD in Computer Science from UC Berkeley, where his advisor was Prof. James Demmel. His research interests include Machine Learning, High-Performance Computing, and Parallel/Distributed Systems. His current work focuses on scaling up deep neural networks training on distributed systems or supercomputers. His team has set world records for ImageNet and BERT training speeds. His LARS and LAMB optimizers are available in MLPerf. He has received several awards, including the Lotfi A. Zadeh Prize, Best Paper Awards at IPDPS and ICPP, and the ACM/IEEE George Michael HPC Fellowship. He is also a Siebel Scholar and was nominated for the ACM Doctoral Dissertation Award.

Yang You is a Presidential Young Professor at the National University of Singapore, specializing in Artificial Intelligence, High Performance Computing, and AI Infrastructure. His research focuses on optimizing deep learning training through large batch optimization, enabling efficient training of models like BERT and ImageNet in significantly reduced time. He has contributed to the development of frameworks such as Colossal-AI, which supports large-scale parallel training, and has explored AI infrastructure for democratizing video production and improving continual learning in vision-language models. His work emphasizes scalability, efficiency, and system-level insights for training long sequences and enhancing model performance.