学术报告(林宙辰 4.24)

卷积神经网络的原则性设计

发布人:杨晓静 发布日期:2021-04-22
主题
卷积神经网络的原则性设计
活动时间
-
活动地址
必赢唯一官方网站315会议室
主讲人
林宙辰 教授 北京大学
主持人
任传贤

The design of convolutional neural networks (CNNs) has undergone two phases: manual design at the early stage, which requires much engineering insights, and the automatic search at the current stage, which heavily relies on computing power. Whether there is an underlying theory for designing good CNNs becomes a crucial research problem. In this talk, I will illustrate our efforts on pursuing this goal. Although I haven’t found a unified principle that can result in all the effective CNNs, I do find multiple principles that can help  design CNNs from various aspects.