ML Systems

Workshop on ML Systems

at NIPS 2017

December 8, 2017

Schedule

08:45 - 09:00     Opening remarks (Sarah Bird)

Session 1 (Chair: Gartn Gibson)

09:00 - 09:20     Invited Talk: Ray: A distributed execution engine for emerging AI applications, Ion Stoica, UC Berkeley

09:20 - 09:40     Contributed Talk: The Case for Learning Database Indexes

09:40 - 10:00     Invited Talk: Federated Multi-Task Learning, Virginia Smith, Stanford University


Session 2: Poster Previews (Chair: Joseph Gonzalez)

10:00 - 10:30     1 min lightning talks

10:30 - 11:30     Posters & Coffee


Session 3 (Chair: Erran Li)

11:30 - 11:50     Invited Talk: Accelerating Persistent Neural Networks at Datacenter Scale, Daniel Lo, Microsoft Research

11:50 - 12:10     Contributed Talk: DLVM: A modern compiler framework for neural network DSLs


12:10 - 13:20     Lunch (Optional Vowpal Wabbit tutorial by John Langford from 12:30 - 13:20, hosted at Extreme Classification workshop)


Session 4: ML Systems Updates (Chair: Aparna Lakshmiratan)

13:20 - 14:50     Updates from Current ML Systems: TensorFlow, PyTorch, Caffe2, CNTK, MXNet, TVM, Clipper, MacroBase, ModelDB

Session 5 (Chair: Yangqing Jia)

14:50 - 15:20     Invited Talk: Machine Learning for Systems and Systems for Machine Learning, Jeff Dean, Google Brain

15:20 - 15:40     Invited Talk: Creating an Open and Flexible ecosystem for AI models with ONNX, Sarah Bird and Dmytro Dzhulgakov, Facebook Research


15:40 - 16:30     Posters & Coffee


Session 6 (Chair: Dan Crankshaw)

16:30 - 16:50     Contributed Talk: NSML: A Machine Learning Platform That Enables You to Focus on Your Models

16:50 - 17:10     Contributed Talk: DAWNBench: An End-to-End Deep Learning Benchmark and Competition

Session 7 (Chair: Siddhartha Sen)

17:10 - 18:15     Panel on Machine Learning Systems Research: Garth Gibson, Joseph Gonzalez, John Langford, Dawn Song, Yangqing Jia