Join us for the twenty-second Google Open Source Live event in our series; “Machine Learning Day”!
Google Machine Learning experts will share updates on everything from Tools for Responsible Machine Learning, to Using Job Queueing to Optimize for Efficiency of Your Compute Resources.
Throughout the event, our speakers will answer selected questions via the Live Q&A Forum. We’ll wrap up the event with an After Party.
Thursday, August 4th from 9:00 am - 11:00 am PST
9:00 Opening for Machine Learning Day on Google Open Source Live
Alexandra Bush, Head of Open Source Marketing and Community (Google Cloud)
Thea Lamkin, Senior Program Manager (Google)
9:03 Session 1: Tools for Responsible Machine Learning
Amy Wang, Product Manager (Google)
Bhaktipriya Radharapu, Software Engineer (Google)
Responsible Artificial Intelligence practices should be embedded within each step of the machine learning journey. We'll introduce an open source toolkit that can help developers ensure model fairness, transparency, and interpretability.
9:25 Session 2: TensorFlow Hub: Machine Learning Models for Everyone
Luiz Gustavo Martins, Artificial Intelligence - Developer Advocate (Google)
Machine Learning can be hard. To build high quality models you’ll need data, expertise and resources to train it. But, is there a better way? Yes, there is! This is where TensorFlow Hub can help. In this talk we will show how you can leverage top researcher’s work to create models suitable for your own needs and by using less resources.
9:47 Session 3: Introduction to On-device Machine Learning
Khanh LeViet, Machine Learning - Developer Advocate (Google)
On-device Machine Learning means running machine learning models on a wide-range of edge devices, such as smartphones or IoT devices. In this session, you’ll learn why on-device machine learning is important, and how to start using it with TensorFlow Lite in just a few lines of code. You’ll also learn about using pre-trained models and training your own custom models.
10:09 Session 4: Using Job Queueing to Optimize for Efficiency of Your Compute Resources
Maciek Różacki, Product Manager (Google)
Organizations are striving for efficiency in their compute environments. Operating with constrained capacity due to physical limits in your data center or commercial in cloud (budget, discounted compute pools, etc.) is something that all infrastructure teams need to deal with. The Batch Working Group in Kubernetes is bringing you new capabilities to cover the need for oversubscribing your resources with a long backlog of Machine Learning jobs that you want to be executed with control, predictability and fairness among the research teams.
10:30 am After Party
11:00 am End
Reasons why you should attend this virtual event LIVE:
- Selected questions will be answered by our speakers in real time! The Live Q&A Forum will be open during the event from 9:00 am to 10:30 am PST.
- Join in on the after party fun, where you can participate in an exciting quiz, and hear from our speakers and emcees immediately following the event!