IDSP-L2: Industry Session on Large-Scale Distributed Learning Strategies |
| Session Type: Lecture |
| Time: Thursday, 7 May, 16:30 - 18:30 |
| Location: On-Demand |
| Virtual Session: View on Virtual Platform |
| Session Chairs: Xiaodong Cui, IBM and Bhuvana Ramabhadran, Google |
| IDSP-L2.1: LOW-RANK GRADIENT APPROXIMATION FOR MEMORY-EFFICIENT ON-DEVICE TRAINING OF DEEP NEURAL NETWORK |
| Mary Gooneratne; Duke University |
| Khe Chai Sim; Google |
| Petr Zadrazil; Google |
| Andreas Kabel; Google |
| Francoise Beaufays; Google |
| Giovanni Motta; Google |
| IDSP-L2.2: IMPROVING EFFICIENCY IN LARGE-SCALE DECENTRALIZED DISTRIBUTED TRAINING |
| Wei Zhang; IBM |
| Xiaodong Cui; IBM |
| Abdullah Kayi; IBM |
| Mingrui Liu; University of Iowa |
| Ulrich Finkler; IBM |
| Brian Kingsbury; IBM |
| George Saon; IBM |
| Youssef Mroueh; IBM |
| Alper Buyuktosunoglu; IBM |
| Payel Das; IBM |
| David Kung; IBM |
| Michael Picheny; IBM |
| IDSP-L2.3: PARALLELIZING ADAM OPTIMIZER WITH BLOCKWISE MODEL-UPDATE FILTERING |
| Kai Chen; Microsoft Research Asia |
| Haisong Ding; University of Science and Technology of China |
| Qiang Huo; Microsoft Research Asia |