Deep Learning Transition Workshop: March 12-13, 2020

** Deadline for applications was January 30, 2020 **


Duke University, Old Chemistry Building – Room 116.


The transition workshop is the capstone of the Deep Learning Program. Each of the working groups presents the work it has done and its members’ plans for future collaboration and research.

Questions: email

Thursday, March 12, 2020
Time Speaker/Talk Slides
8:30 Registration and Welcome
Topic: Bayesian Methods in Deep Learning
9:00-9:30 David Dunson, Duke University
9:30-10:00 Deborshee Sen, SAMSI
Bayesian Dimension Reduction using Neural Networks
10:00-10:30 Bianca Dumitrascu, SAMSI
10:30-11:00 Break/Conversations
Topic: Interpretable Deep Networks
11:00-11:30 Cynthia Rudin, Duke University
11:30-Noon Haiyang Huang, Duke University
Dimension Reduction and Manifold Learning: a survey
Noon-12:30 TBD
12:30-2:00 Lunch on your own
2:00-2:30 Pulong Ma, SAMSI
Kriging: Beyond Matern
2:30-3:00 Anindya Bhadra, Purdue University
Deep Neural Network Emulators: beyond Gaussian Processes
3:00-3:30 Xiyuan Cheng, Duke University
3:30-4:00 Break
Topic: Regularization Techniques for Training Deep Networks
4:00-4:30 Sorin Mitran, University of North Carolina
4:30-5:00 Quoc Tran-Dinh, University of North Carolina
Shuffling and Sample-Based Schemes for Non-Convex Optimization
5:00-5:30 Linjun Zhang, Rutgers University
Exploring Model Sensitivity via Adversarial Influence Functions
5:30-7:00 Poster Session and Reception

Friday, March 13, 2020

Time Speaker/Talk Slides
Topic: Miscellany
9:00-10:00 Jhuma Das, University of North Carolina; Adrian Green, North Carolina State University; Martin Mohlenkamp, Ohio University
Leveraging High-Throughput Screening Data
10:00-10:30 David Banks, SAMSI
Teaching Deep Learning
10:30-11:00 Break
11:00-11:30 Guang Cheng, Purdue University
Classification under Teacher-Student Network: sharp rate of convergence
11:30-12:30 Review of Working Group Best Practices
12:30 Adjourn