Machine Learning is becoming increasingly pervasive and AWS has a plethora of services from which to choose. This guide provides a comprehensive view of ML services, including newly released ML APIs, and by attending a selection of these events, you’ll learn how to start to use machine learning in the AWS Cloud. This guide is designed for developers who are looking to build applications, which use AWS machine learning services. You’ll notice that I’ve included a number of sessions which focus on SageMaker – it’s a key new service to master in this area.
- AIM301-R - Deep Learning for Developers: An Introduction
Learn key concepts of deep learning, such as neural networks, activation functions and optimizers, so that you can understand what types of business problems deep learning fits best. Includes examples of implementing complex pattern recognition models for pictures, text and sounds. Get an intro to the most commonly used DNNs, TensorFlow and MXNet.
- AIM404-R - Build, Train, and Deploy ML Models Quickly and Easily with Amazon SageMaker, ft. 21st Century Fox
Understand how SageMaker supports fast ML model training and deployment in this practical session. See customer examples of using the various feature sets of SageMaker to deliver ML to your customers faster.
- AIM302 - Machine Learning at the Edge
See how to use Greengrass to deploy machine learning to edge devices, which may or may not be connected to the cloud. This functionality is useful for IoT scenarios, such as oil and gas rig monitoring, along with newer solutions, such as pattern recognition via devices (cameras, etc.…). You’ll see how SageMaker and Greengrass can work together in these advanced scenarios.
- Stop by the Welcome Reception in The Venetian
- Watch "Monday Night Live" from 7:30–9PM in The Venetian
- Run the 4K or 8K race…or consider donating to Girls Who Code (proceeds from the run support it).
- AIM406 - Unlock the Full Potential of Your Media Assets
Use ML APIs (Rekognition, Translate and Comprehend) to quickly analyze video assets. These APIs extract metadata automatically and include functionality such as language translation and face (include celebrity face!) recognition. These tools help you build searchable video libraries, automate creation of highlight reels, and other types of categorization.
- ANT318-R - Build, Deploy and Serve Machine learning models on streaming data using Amazon Sagemaker, Apache Spark on Amazon EMR and Amazon Kinesis
As data exponentially grows in organizations, there is an increasing need to use machine learning (ML) to see a powerful end-to-end ML workflow in action. Here you’ll understand how to scale a complex workflow via integration with open source libraries (Apache Spark) and AWS services. Specifically, you’ll see EMR, Kinesis and SageMaker ‘talking to each other’, all of which can enable scenarios such as massive near-real time prediction pipelines.
- AIM402-R - Deep Learning Applications Using PyTorch
PyTorch is often used by researchers to build deep neural networks. See how to build a CNN with PyTorch and SageMaker in this deep session (which includes lots of code examples).
- Enjoy the Pub Crawl in any number of the great locations!
- ENT232-R - Machine Learning for the Enterprise, ft. Sony Interactive Entertainment
A practical example of using SageMaker to rapidly build powerful and flexible ML models which add value to your applications. Here customer Sony Interactive details how they used SageMaker, along with AWS ML APIs for vision and language to implement ML for the enterprise.
- Transform the Modern Contact Center Using Machine Learning and Analytics
Here’s another practical example showing how to use ML services to improve service for a common use case – a contact center. Again, the concept is using ML to serve more customers faster. Specifically, you’ll see how to use speech and text analytics to be alerted to potentially problematics service-related trends faster.
- AIM401-R - Deep Learning Applications Using TensorFlow
TensorFlow is the most widely used library for building custom, complex deep learning models. Understand when and how to use it is key to building effective machine learning solutions. Also see how TF integrates with SageMaker so you can deploy your models fasters.
- AIM405-R - Better Analytics Through Natural Language Processing
Mixing old and new is smart business. Here you’ll see how to work with data in RDS or Elasticsearch using newer services, such as Neptune (graph database) and also ML APIs for text analytics. Specifically, you’ll work with Comprehend for text analytics for natural language processing.
- AIM303-R3 - Create Smart and Interactive Apps with Intelligent Language Services on AWS
I am obviously keen on ML solutions for natural language – and here’s another great session on this topic. Chatbots are here – see how to use AWS services to quickly build a chatbot that adds value for you and your customers.
- AIM403-R - Integrate Amazon SageMaker with Apache Spark, ft. Moody's
Apache Spark and SageMaker equal powerful complex analytics at scale – who doesn’t want that? SageMaker includes a number of commonly used ML algorithms that have been optimized to work with AWS services, learning how this works will save you time and money.
- Stop by the re:Play party (runs from 8PM to midnight)
Enjoy these ML-focused events. As a Cloud Architect and Developer, I'm excited by the possibilities presented by the AWS Cloud. Understanding which machine learning service to use in your solution is a key solution differentiator for application developers and architects.
For a comprehensive guide to the rest of the conference, check out AWS Community Hero Mark Nunnikhoven’s post “The Ultimate Guide to AWS re:Invent 2018”. Also I’ll be live tweeting during the conference - where I’m @lynnlangit.