Schedule

LIVE STREAM AVAILABLE

09:00 - 09:10     Opening remarks (Sarah Bird)

Session 1: ML Programmability (Chair: Sarah Bird)

09:10 - 09:40     Keynote 1: Accelerated Computing for AI, Bryan Catanzaro, NVIDIA: Slides

09:40 - 10:00     Contributed Talk: Fashionable Modeling with Flux, Mike J. Innes, Julia Computing

10:00 - 10:20     Contributed Talk: Model Assertions for Debugging Machine Learning, Daniel Kang and Deepti Raghavan, Stanford University

Session 2: Morning Poster Session (Chair: Siddhartha Sen)

10:20 - 10:35     Poster Introduction: Siddhartha Sen

10:35 - 11:40     Posters & Coffee

Session 3: Large-scale ML (Chair: Aparna Lakshmi Ratan)

11:40 - 12:10     Keynote 2: BREAKING NEWS: Types Found to Be Underrated, Aish Fenton, Netflix

12:10 - 12:30     Contributed Talk: Parallel Training of Linear Models Without Compromising Convergence, Nikolas Ioannou, IBM Research

Session 4: Open Source Software Showcase with Catered Lunch (Chair: Dan Crankshaw)

12:30 - 12:45     Pick up lunch. The workshop will be providing catered lunch sponsored by Facebook and Microsoft.

12:50 - 02:55     Open Source Software Showcase (10 minutes each)

• TensorFlow - Rajat Monga, Google
• PyTorch - Soumith Chintala, Facebook
• Keras - François Chollet, Google
• MXNet - Alex Smola, Amazon
• Chainer - Seiya Tokui, Preferred Networks
• Ray - Robert Nishihara, University of California, Berkeley
• TVM and VTA - Tianqi Chen and Thierry Moreau, University of Washington
• CoreML - Gaurav Kapoor, Apple
• ML.NET - Markus Weimer, Microsoft
• MLFlow - Matei Zaharia, Databricks and Stanford

Session 5: Afternoon Poster Session

02:55 - 03:40     Posters & Coffee

Session 6: Rethinking ML Infrastructure (Chair: Aparna Lakshmi Ratan)

03:40 - 04:10     Keynote 3: Infrastructure and Systems for Applied Machine Learning at Facebook, Kim Hazelwood, Facebook

04:10 - 04:30     Contributed Talk: Accelerating Deep Learning Workloads Through Efficient Multi-Model Execution, Deepak Narayanan, Stanford

04:30 - 04:50     Contributed Talk: Rethinking Floating Point for Deep Learning, Jeff Johnson, Facebook AI Research

04:50 - 05:10     Contributed Talk: A Case for Serverless Machine Learning, Joao Carreira, UC Berkeley

05:10 - 05:30     Closing remarks (Aparna Lakshmi Ratan)