19 RSVP'd
Welcome to TensorFlow Everywhere - North America!
This is part of a global community-led event series for TensorFlow and Machine Learning enthusiasts and developers around the world, powered by the North America Developer Ecosystem at Google.
Join us on 26 & 27 Feb 2021 via YouTube
You will hear from the TensorFlow team and Machine Learning Google Developer Experts on latest updates!
Registration page: https://events.withgoogle.com/tensorflow-everywhere-na/
Tentative Schedule in Pacific Standard Time
26 Feb 2021 (Fri)
9.00 - 9.10 AM PST - Welcoming Remarks & Overview of Google Developers Program in North America
Kyle Paul - Google Developer Ecosystem, North America Regional Lead
9.10 - 9.40 AM PST - What’s New in TensorFlow 2.4?
Laurence Moroney - Google Developer Advocate
A technical round up of what's new in TF in 2020, including TF Agents for reinforcement learning, tf.experimental.numpy, and tools for exploring your model's performance. Join us for a rapid-fire look at many new features.
9.40 AM - 10.00 AM PST - Responsible ML in Production
Catherine Nelson - Google Developer Expert Machine Learning
When your machine learning model is deployed to a production system, this is a critical time: your model starts to interact with real people. This is the perfect moment to check that your model’s predictions aren’t showing any harmful biases. In this lightning talk, I will introduce a production pipeline using the TensorFlow ecosystem that includes ways to identify and reduce harmful impacts.
10.00 - 10.45 AM PST - Intro to Deep Learning
Josh Gordon - Google Developer Advocate
A quick intro to deep learning, including representation learning, and fundamentals of neural networks. You'll learn about the major types of neural networks (DNNs, CNNs, and RNNs), as well as techniques that can help you interpret how and why they work, including Deep Dream. Each area mentioned will include links to complete code examples in TensorFlow 2.0.
10.45 - 11.25 AM PST - Predicting the S&P500 with Recurrent Neural Networks
Ekaba Bisong - Google Developer Expert Machine Learning
Recurrent neural networks are a special type of deep neural networks for learning outcomes where future predictions are linked to past events. The S&P 500 is a stock market index that tracks the performance of the 500 largest companies that are listed on stock exchanges in the United States. We may observe that the future price (in this case closing price) of an underlying (in this case, the S&P) may be linked to the past market data such as the previous high, low, opening and closing price of the underlying and the volume of trades made per day. Other information that affect the market in the past may also be contributing factors. In this project, we frame predicting the future closing price of the S&P 500 as a sequence problem. Hence, we use recurrent neural networks to build an approximation model for a multivariate time-series problem.
11.25 AM - 12.05 PM PST - Machine Learning-powered Pipelines to Augment Human Specialists
Tanmay Bakshi - Google Developer Expert Machine Learning
Machine Learning techniques are particularly helpful for analyzing large datasets to help humans find insights at scale. In the field of biology, scientists without expensive equipment spend hours doing repetitive tasks that add no value to their research, but are required to run experiments and to come to conclusions, like cell counting. In this talk, let's dive into how a deep learning powered pipeline using TensorFlow enables researchers to come to conclusions from data faster using inexpensive equipment, like the cameras on mobile phones.
12.05 - 12.50 PM PST - ML Engineering at Digits
Hannes Hapke - Google Developer Expert Machine Learning
12.50 - 1.00 PM PST
Kyle Paul - Google Developer Ecosystem, North America Regional Lead
Day 1 Wrap Up
27 Feb 2021 (Sat)
9.00 - 9.10 AM PST - Welcoming Remarks
Kyle Paul - Google Developer Ecosystem, North America Regional Lead
9.10 - 10.00 AM PST - Be Bach on a Budget
Vikram Tiwari - Google Developer Expert Google Cloud Platform, Machine Learning
Have you always wanted to get creative but lacked the talent? With machine learning, you too can become creative! We will explore how to start using machine learning to solve every day challenges and how to go beyond into the avenues of creative exploration with machines helping you. In this talk we cover TensorFlow.js with Project Magenta, various machine learning techniques like style transfer, RNNs, AutoEncoders etc.
10.00 - 10.45 AM PST - Superpowers for next gen web apps: Machine learning
Jason Mayes - Google Senior Developer Relations Engineer for TensorFlow.js
Discover how to achieve superpowers by embracing machine learning in JavaScript using TensorFlow.js in the browser. Learn what machine learning is, get a high-level overview of how it works, get inspired through a whole bunch of creative prototypes (from invisibility to teleportation) that push the boundaries of what is possible in the modern web browser, and then take your own first steps with machine learning in minutes. By the end of the session, everyone (no matter what your background) will understand how to recognize an object of their choice, which could then be used in any creative way you can imagine on your own website. No background in machine learning is required. Take your first steps with TensorFlow.js!
10.45 - 11.25 AM PST - MLOps at the Edge
Leigh Johnson - Google Developer Expert Machine Learning
Operationalizing machine learning is tough enough in your own datacenter or virtual private cloud. When the target is a low-powered IoT device, the versioning, deployment, and monitoring of ML models comes with additional challenges. In this talk, I'll show you how I developed a defect detection and quality monitoring system for 3D printers. I'll describe the transport (MQTT) and mechanisms for continuous retraining and redeployment of TensorFlow models to Raspberry Pi, using Cloud IoT Core and DataFlow (Apache Beam). I'll cover the trade-offs required for real-time inference performed entirely on-device with TensorFlow Lite, plus the techniques I used to study the performance impact of model quantization. Key concepts/technologies: Edge ML, TensorFlow Lite, TensorFlow Object Detection API, DataFlow (Apache Beam), Cloud IoT Core.
11.25 AM - 12.05 PM PST - Zero to Hero with TensorFlow (Getting Started for Developers)
Laurence Moroney - Google Developer Advocate
If you're interested in ML, but didn't know where to start, this talk will help you understand the paradigm of ML and how TensorFlow can be used to train machine-learned models. You'll see how to use some of the latest and greatest features in TensorFlow that let you go from nothing to a working ML model in just a few minutes. No PhD required.
12.05 - 12.50 PM PST - AI from Magic to Science
Gant Laborde - Google Developer Expert Machine Learning, Web Technologies
TensorFlow.js is a fantastic opportunity to level up everyone's understanding of machine learning. What was considered magic a decade ago has become the science of today. So what's the science of tomorrow? Let's construct and build a fun and exciting example of TensorFlow.js that excites and invigorates TensorFlow everywhere!
12.50 - 1.35 PM PST
TBC
12.50 - 1.00 PM PST - Closing Remarks
Kyle Paul - Google Developer Ecosystem, North America Regional Lead
Contact Us