AWS re:Invent 2018

Deploying deep learning applications at scale can be cost prohibitive due to the need for hardware acceleration to meet latency and throughput requirements of inference. Amazon Elastic Inference helps you tackle this problem by reducing the cost of inference by up to 75% with GPU-powered acceleration that can be right-sized to your application's inference needs. In this session, learn about how to deploy TensorFlow, Apache MXNet, and ONNX models with Amazon Elastic Inference on Amazon EC2 and Amazon SageMaker. Hear from Autodesk on the positive impact of AI on tools used to design and make a better world. Learn about how Autodesk and the Autodesk AI Lab are using Amazon Elastic Inference to make it cost efficient to run these tools at scale.

Developers start your engines! This breakout session will provide an introduction to the newly launched AWS DeepRacer. Learn about the basics of reinforcement learning, what's under the hood and your opportunities to experience AWS DeepRacer for yourself.

Successful machine learning models are built on high-quality training datasets. Labeling raw data to get accurate training datasets involves a lot of time and effort because sophisticated models can require thousands of labeled examples to learn from, before they can produce good results. Typically, the task of labeling is distributed across a large number of humans, adding significant overhead and cost. Join us as we introduce Amazon SageMaker Ground Truth, a new service that provides an effective solution to reduce this cost and complexity using a machine learning technique called active learning. Active learning reduces the time and manual effort required to do data labeling, by continuously training machine learning algorithms based on labels from humans. By iterating through ambiguous data points, Ground Truth improves the ability to automatically label data resulting in high-quality training datasets.

Until now, many customers spent time creating or searching for the right algorithm and model when using Amazon SageMaker. In this session you'll learn how to build machine learning applications even faster by finding curated algorithms and model packages in AWS Marketplace and deploying them directly on Amazon SageMaker.

Today, organizations deploy more AI/ML workloads on AWS than on any other cloud platform. The cloud has removed many of the challenges associated with scalability, and it's never been easier or more cost effective to build custom and intelligent data models. In this session, learn how the C3 Platform leverages the full power of Intel Xeon Scalable processors on AWS to rapidly train, deploy, and operationalize AI/ML and big data applications like C3 Inventory Optimization and C3 Predictive Maintenance. In addition, a customer shares how these solutions helped achieve demonstrable value. This session is brought to you by AWS partner, Intel.

Come see examples of how Bebo uses Amazon SageMaker to power massive Fortnite tournaments every week. Traditional sports require referees, scorekeepers, field staff, and broadcast crews for every match. But esports are digital by nature. In this session, learn how machine learning and computer vision are enabling esports to occur at a massive scale. Learn how Bebo developed a model that can detect every victory and elimination, and can even prevent cheating on their tournament platform.

In this session, we cover best practices for enterprises that want to use powerful open-source technologies to simplify and scale their machine learning (ML) efforts. Learn how to use Apache Spark, the data processing and analytics engine commonly used at enterprises today, for data preparation as it unifies data at massive scale across various sources. We train models using TensorFlow, and we use MLflow to track experiment runs between multiple users within a reproducible environment. We then manage the deployment of models to production. We show you how MLflow can be used with any existing ML library and incrementally incorporated into an existing ML development process. This session is brought to you by AWS partner, Databricks.

The TensorFlow deep learning framework is used for developing diverse artificial intelligence (AI) applications, including computer vision, natural language, speech, and translation. In this session, learn how to use TensorFlow within the Amazon SageMaker machine learning platform. Then, hear from Advanced Microgrid Solutions about how they implemented a deep neural network architecture with Keras and TensorFlow to forecast energy prices in near real time. Complete Title: AWS re:Invent 2018: [REPEAT 2] Deep Learning Applications Using TensorFlow, ft. Advanced Microgrid Solutions (AIM401-R2)

With support for PyTorch 1.0 on Amazon SageMaker, you now have a flexible deep learning framework combined with a fully managed machine learning platform to transition seamlessly from research prototyping to production deployment. In this session, learn how to develop with PyTorch 1.0 within Amazon SageMaker using a novel generative adversarial network (GAN) tutorial. Then, hear from Facebook on how you can use the FAIRSeq modeling toolkit, which serves 6B translations daily for Facebook users, to train your own custom PyTorch models on Amazon SageMaker. Facebook also discusses the evolution of PyTorch 1.0 and features introduced to accelerate research and deployment. Complete Title: AWS re:Invent 2018: [REPEAT 1] Deep Learning Applications Using PyTorch, Featuring Facebook (AIM402-R1)

Amazon SageMaker, our fully managed machine learning platform, comes with pre-built algorithms and popular deep learning frameworks. Amazon SageMaker also includes an Apache Spark library that you can use to easily train models from your Spark clusters. In this code-level session, we show you how to integrate your Apache Spark application with Amazon SageMaker. We also dive deep into starting training jobs from Spark, integrating training jobs in Spark pipelines, and more.