学术报告（叶荫宇 2026.1.5）

摘要：This talk aims to present several mathematical optimization problems, models, and algorithms for AI such as the LLM training and inferencing. In particular, we describe how classic optimization models/theories, such as online linear programming, resource allocation and gradient-conditioning, may be applicable to accelerate and improve the Training/Inferencing processes that used in LLMs. We also discuss the recent progresses of classical Optimization Solvers and their applications for sciences and industries.