报告题目：From deep learning to federated learning and everything in between

报告人：美国西北大学深度学习中心主任Diego Klabjan教授

报告时间：2022年11月3日（周四）晚上21:00点

报告地点：腾讯会议线上（会议号：940 336 429；密码：221103）

主持人：李天瑞教授、罗志鹏助理教授

报告摘要：

Time series problems exhibit many different feature types that should be carefully considered in deep learning recurrent models. Regarding this input side challenge, we discuss how to tailor recurrent cells. On the output side, one of the most vaxing questions is how far in the future to predict. Our work suggests loss functions to cope with this. There is a growing trend in developing models that require few data samples. We address this question by domain adaptation in an open set setting. All these models require full access to data while privacy advocates are becoming loud. Hybrid federated learning is tricky to handle, and facing with non-iid data even more dauting. We discuss algorithms that address such situations.

报告人简介：

Diego Klabjan is a professor at Northwestern University, Department of Industrial Engineering and Management Sciences. He is also Founding Director, Master of Science in Analytics and Director, Center for Deep Learning. His research is focused on machine learning, deep learning and analytics (modeling, methodologies, theoretical results) with concentration in finance, insurance, sports, and bioinformatics. Professor Klabjan has led projects with large companies such as Intel, Baxter, Allstate, Anthem, and many others. He is also a co-founder of Videspan, LLC after successfully starting Opex Analytics, a Coupa company.