WA4a: Deep Learning Theory (Invited) |
| Session Type: Oral |
| Time: Wednesday, November 6, 08:15 - 09:55 |
| Location: Heather |
| Session Chair: Tom Goldstein, University of Maryland, College Park |
| WA4a-1: A THEORETICAL LOOK AT ADVERSARIAL ATTACKS ON NEURAL NETWORKS |
| Tom Goldstein; University of Maryland |
| Ali Shafahi; University of Maryland |
| W Ronny Huang; University of Maryland |
| Soheil Feizi; University of Maryland |
| WA4a-2: RANDOMIZED ALTERNATING DIRECTIONS METHODS FOR EFFICIENT DISTRIBUTED OPTIMIZATION |
| Mert Pilanci; Stanford University |
| WA4a-3: ROBUSTNESS AND GENERALIZATION GUARANTEES FOR OVER-PARAMETERIZED NEURAL NETWORK TRAINING |
| Mahdi Soltanolkotabi; University of Southern California |
| WA4a-4: LOW-DIMENSIONAL MODELS AND DEEP NETWORKS FOR HIGH-DIMENSIONAL DATA |
| Yi Ma; University of California, Berkeley |