So you've trained a cool machine learning model - now what?

Nov 13, 2018 · Mountain View, United States of America

Hello Makers!

Join us this evening to hear from makers behind H2O Driverless AI and Technical Staff of Oracle! following is a brief agenda for the evening:

6 - 6:30 - Doors open and networking
6:30 - 7:30 - Anthony and Mark's talk
7:30 - 8 - Q&A

Talk 1: AlphaZero on GraphPipe
In this introductory discussion, X from Oracle Cloud Infrastructure will walk through the essential elements of taking neural network models from R&D to production. The discussion will include a survey of prominent model formats, including Tensorflow, Caffe2, ONNX, and TensorRT, and discuss how one can deploy these models to production using various serving technologies like GraphPipe.

AlphaZero on GraphPipe - Accelerated training of the AlphaZero algorithm using GraphPipe AlphaZero is an interesting ML case study, as it requires massive amounts of model inference for its game generation phase. In this talk, we discuss the challenges and bottlenecks of the AlphaZero algorithm, and go into detail about how we used GraphPipe at a variety of key points in our architecture to overcome them, from initial AlphaZero training all the way through front-end web-based application deployment.

Talk 2: Interpretable Machine Learning
The good news is building fair, accountable, and transparent machine learning systems is possible. The bad news is it’s harder than many blogs and software package docs would have you believe. The truth is nearly all interpretable machine learning techniques generate approximate explanations, that the fields of eXplainable AI (XAI) and Fairness, Accountability, and Transparency in Machine Learning (FAT/ML) are very new, and that few best practices have been widely agreed upon. This combination can lead to some ugly outcomes!

This talk aims to make your interpretable machine learning project a success by describing fundamental technical challenges you will face in building an interpretable machine learning system, defining the real-world value proposition of approximate explanations for exact models, and then outlining the following viable techniques for debugging, explaining, and testing machine learning models

Speaker Bios:
Speaker 1
Anthony is a Consulting Member of Technical Staff at Oracle. He works in OCI where he focuses on the intersection of distributed computing and AI. Previous to Oracle, he was Principal Engineer at Whitepages, where he architected data products, created ML-driven fraud detection infrastructure, and initiated their migration from Data Center to cloud. Anthony was also on the founding team of OpenStack, and worked as a core developer on Cinder, Horizon, and Devstack.

Speaker 2
Mark is a hacker at H2O. He was previously in the finance world as a quantitative research developer at Thomson Reuters and Nipun Capital. He also worked as a data scientist at an IoT startup, where he built a web based machine learning platform and developed predictive models.
Mark has a MS Financial Engineering from UCLA and a BS Computer Engineering from University of Illinois Urbana-Champaign. In his spare time Mark likes competing on Kaggle and cycling.

Event organizers
  • H2O AI & Deep Learning

    Welcome to the group. We’re excited to bring you the latest happenings in AI, Machine Learning, Deep Learning, Data Science and Big Data. Who are we? We’re, creators of the world’s leading open source deep learning and machine learning platform, used by more than 90,000 data scientists and 9,000 organizations around the world. Our goal is to congregate with data enthusiasts discuss trending topics in the world of AI. We also regularly invite esteemed industry influencers and thought leaders who ta

    Recent Events

Are you organizing So you've trained a cool machine learning model - now what??

Claim the event and start manage its content.

I am the organizer

based on 0 reviews