报告题目：Continuous-depth neural networks
报告时间：2022/12/01 09:00-10:00 
报告地点：#腾讯会议：260-442-410（//meeting.tencent.com/dm/Eu1X1PskCHRG）
报告摘要：A series of recent works have revealed a close connection between neural networks and dynamical systems. In this talk, I will give a brief introduction to the continuous-depth neural networks, such as the neural ordinary differential equations (Neural ODEs). In the last five years, the Neural ODEs have been widely applied, showing exceptional efficacy in coping with some representative datasets. In the framework of Neural ODEs, it parameterizes the vector filed of ODEs by a neural network. The output (terminal state) of Neural ODEs is computed via a black-box ODE solver. Last, I will introduce my own work on this filed, such as the neural delay differential equations (Neural DDEs), a variant of Neural ODEs.
报告人：朱群喜，复旦大学智能复杂体系基础理论与关键技术实验室
报告人简介：主要从事复杂网络和深度学习的研究工作, 相关成果发表于国际顶级期刊IEEE TAC, SICON, SCL, CHAOS和人工智能顶级会议ICLR 2021, 得到同行大家的评价引用 (Edward Ott、Jürgen Kurths、曹进德、桂卫华等院士, Google Research和UC Berkeley等团队）。
邀请人：马欢飞