Seminar Announcement

Deep Learning for Pixel-Level Predictive and Generative Modeling

  • Speaker: Prof. Shuiwang Ji
  • Washington State University
  • Date: Wednesday, Nov. 1, 2017
  • Time: 1:30pm - 2:30pm
  • Location: Room VTRC-A: East Falls Church Room

Abstract

Deep learning methods take raw signals as inputs and compute hierarchical representations automatically. For example, traditional convolutional neural networks use convolutional and pooling layers to compute features of decreasing spatial/temporal sizes in classification tasks. Recent deep models for pixel-level predictive and generative modeling (including VAE, GAN, and U-Net) require generating features of increasing spatial/temporal sizes. In this talk, I will discuss a few limitations and challenges of current methods for pixel-level predictive and generative modeling. I will then present our work on addressing these challenges using principled, fundamental techniques. I will present case studies on how to use our methods for solving challenging problems in neuroscience, biology, and medicine.

Speaker's Biography

Shuiwang Ji is an Associate Professor in the School of Electrical Engineering and Computer Science at Washington State University. He received the Ph.D. degree in Computer Science from Arizona State University in 2010. His research interests include machine learning, data mining, computational biology, and brain data analytics. Shuiwang Ji received the National Science Foundation CAREER Award in 2014. Currently, he serves as an Associate Editor for ACM Transactions on Knowledge Discovery from Data, IEEE Transactions on Neural Networks and Learning Systems, and BMC Bioinformatics. He was a program chair for the 2017 Bioimage Informatics conference and serves as a senior program committee member for IJCAI, KDD, and SDM. He has served as a technical program committee member of major conferences in machine learning (ICML, NIPS), data mining (KDD, SDM, ICDM), and bioinformatics and medical image computing (MICCAI and PSB).