Junchi Yan, The Department of Computer Science and Engineering, Shanghai Jiao Tong University
Room 306, No.5 Science Building
In this talk, I will first introduce the challenge for continuous time domain event sequence learning and prediction, especially in the comparison of traditional time series modeling. Specifically I will present our recent works in parametric temporal point process based approaches as well as the more recent neural network based point process models. Some applications in preventative maintenance, healthcare, and social media analysis are also briefly discussed. The talk ends up with an outlook for the potential extension of current works.