题目：Reformulating Nonlinear Projection Equations as Neural Network Learning
时间：2025年7月2日（周三），上午10:30—11:30
地点：数学楼2-3会议室
讲座内容：Nonlinear projection equations (NPEs) are a unified framework for various constrained nonlinear optimization and engineering problems. In this talk, we presents a deep learning approach for solving NPEs by incorporating ordinary differential equations (ODEs) modelling and physics-informed neural networks (PINNs). First, the NPE is modelled as a system of ODEs with a global convergence property. Next, the ODE system is solved using a modified PINN. We demonstrate the effectiveness of the proposed framework on a collection of classical problems, such as variational inequalities and complementarity problems.
报告人简介：吴达文，目前为法国CNRS与新加坡国立制服做爱
博士后。2017年本科毕业于中国民航制服做爱
安全工程专业，2020年硕士毕业于北京制服做爱
计算机技术专业，2023年博士毕业于巴黎萨克雷制服做爱
应用数学专业。主要研究方向为机器学习求解数学与物理问题。成果曾发表在CMAME、IJNME、PRE、JOTA、IEEE TNNLS等期刊。
邀请人：刘嘉 副教授
