Shuicheng Yan
Member of Singapore Academy of Engineering, IEEE Fellow, Honourary Advisor of Kunlun Wanwei
Prof Yen Shui Seng is currently the Director of KunlunWanwei 2050 Research Institute, Chief Scientist of KunlunWanwei and Skywork AI, and was formerly the Chief Scientist of Sea. He is a member of the Singapore Academy of Engineering and a Fellow of AAAI, ACM, IEEE and IAPR, and his research interests include computer vision, machine learning and multimedia analytics. To date, Prof Yen Shui-Shing has published more than 800 papers in top international journals and conferences with an H-index of 140+. Prof Yen Shuicheng's team has won more than ten awards in two core competitions, Pascal VOC and ImageNet (ILSVRC). In addition, his team has won more than ten best paper and best student paper awards, especially a grand slam at ACM Multimedia, the top conference in multimedia, including three best paper awards, two best student paper awards, and the best presentation award.
Topic
Foundations of Foundation Models
In this sharing, I divide the research on foundation models into three aspects, corresponding to three errors, approximation eror, estimation error and optimization error. For approximation error, I shall share two recent new progresses on model architecture, MoE++ and MoH. For estimation error, I shall share our progresses on how data curation helps reward model training. For optimization error, I shall share our new optimizer called Adan and the work on how to optimize super-parameters. At the same time, I shall share two of our new products, SkyMusic and SkyReels.
