AWS re:Invent 2017

In machine learning, training large models on massive amount of data usually improved results. Our customers report, however, that training such models and deploying them is either operationally prohibitive or outright impossible for them. Amazon AI Algorithms is designed to solve this problem. It is a collection of distributed streaming ML algorithms that scale to any amount of data. They are fast and efficient because they distribute across CPU/GPU machines and share a collective distributed state via a highly-optimized parameter server. They scale to an infinite amount of data because they operate in the streaming model. This means they require only one pass over the data and never increase their resources consumption, allowing training to be paused, resumed, and snapshotted and even for algorithms to consume kinesis streams directly providing an “always on” training mechanism.  They are production ready.  Trained models are automatically containerized and useable in production using Amazon SageMaker hosting. Finally, we provide a convenient SDK which allows scientists to create new algorithms which operate in this model and enjoy all the benefits above. This talk will discuss our design choices and some of the internal working of the system. It will also describe the distributed streaming model and its numerous benefits to machine learning practitioners. We will show how to invoke large scale learning from Amazon SageMaker, or Amazon EMR, and host the solution. Time permits, we will show how to develop a new Algorithm using the SDK.

Customers have several options of architecting recommendation engines, and a graphdb is the best way to create a real-time recommendation engine. I will use the as-yet unreleased Neptune service and will show a demo.In this session, we will look at approaches to use machine learning and graph representations for Cyber Investigative Analytics. We will give a demonstration of Graphistry using Amazon Neptune and a graph-based approach to detecting anomalies in Netflow data.

The need for Natural Language Processing (NLP) is gaining more importance as the amount of unstructured text data doubles every 18 months and customers are looking to extend their existing analytics workloads to include natural language capabilities. Historically, this data had been prohibitively expensive to store and early manual processing evolved into rule-based systems, which were expensive to operate and inflexible.  In this session we will show you how you can address this problem using Amazon Comprehend.

Building a conversational AI experience that can respond to a wide variety of inputs and situations depends on gathering high-quality, relevant training data. Dialog with humans is an important part of this training process. In this session, learn how researchers at Facebook use Amazon Mechanical Turk within the ParlAI (pronounced “parlay”) framework for training and evaluating AI models to perform data collection, human training, and human evaluation. Learn how you can use this interface to gather high-quality training data to build next-generation chatbots and conversational agents.

Ever since the term “crowdsourcing” was coined in 2006, it's been a buzzword for technology companies and social institutions. In the technology sector, crowdsourcing is instrumental for verifying machine learning algorithms, which, in turn, improves the user's experience. In this session, we explore how Pinterest adapted to an increased reliability on human evaluation to improve their product, with a focus on how they've integrated with Mechanical Turk's platform. This presentation is aimed at engineers, analysts, program managers, and product managers who are interested in how companies rely on Mechanical Turk's human evaluation platform to better understand content and improve machine learning algorithms. The discussion focuses on the analysis and product decisions related to building a high quality crowdsourcing system that takes advantage of Mechanical Turk's powerful worker community.

Artificial intelligence is going to be part of every software workload in the not-too-distant future. Partnering with AWS, Intel is dedicated to bringing the best full-stack solutions to help solve business and societal problems by helping turn massive datasets into information. Thorn is a non-profit organization, co-founded by Ashton Kutcher, focused on using technology innovation to combat child sexual exploitation. It is using MemSQL to provide a new approach to machine learning and real-time image recognition by making use of the high-performance Intel SIMD vector dot product functionality. This session covers machine learning on Intel Xeon processor based platforms and features speakers from Intel, Thorn, and MemSQL. Session sponsored by Intel

In this talk, you will learn how to use, or create Deep Learning architectures for Image Recognition and other neural network computations in Apache Spark. Alex, Tim and Sujee will begin with an introduction to Deep Learning using BigDL. Then they will explain and demonstrate how image recognition works using step by step diagrams, and code which will give you a fundamental understanding of how you can perform image recognition tasks within Apache Spark. Then, they will give a quick overview of how to perform image recognition on a much larger dataset using the Inception architecture. BigDL was created specifically for Spark and takes advantage of Spark's ability to distribute data processing workloads across many nodes. As an attendee in this session, you will learn how to run the demos on your laptop, on your own cluster, or use the BigDL AMI in the AWS Marketplace. Either way, you walk away with a much better understanding of how to run deep learning workloads using Apache Spark with BigDL. Session sponsored by Intel

Amazon SageMaker is a fully-managed service that enables data scientists and developers to quickly and easily build, train, and deploy machine learning models, at scale. This session will introduce you the features of Amazon SageMaker, including a one-click training environment, highly-optimized machine learning algorithms with built-in model tuning, and deployment without engineering effort.  With zero-setup required, Amazon SageMaker significantly decreases your training time and overall cost of building production machine learning systems.  You'll also hear how and why Intuit is using Amazon SaeMaker on AWS for real-time fraud detection.

In this session, you'll learn how to leverage AWS Service Catalog, AWS Lambda, AWS Config and AWS CloudFormation to create a robust, agile environment while maintaining enterprise standards, controls and workflows. Fannie Mae demonstrates how they are leveraging this solution to integrate with their existing workflows and CMDB/ITSM systems to create an end-to-end automated and agile IT lifecycle and workflow.

Organizations use application delivery controllers (ADCs) to ensure that their most important applications receive the best performance across their network. In this session, you learn how and why Salesforce used the F5 BIG-IP platform, an ADC solution from AWS Marketplace, during a migration to AWS. To preserve an existing skillset within their business, Salesforce chose AWS Marketplace to first evaluate the solution on the AWS platform before ultimately selecting it as part of their international rollout. You see how BIG-IP performs application routing and security, and how it works with existing AWS networking solutions to provide a consistent experience for domestic and international rollouts. You also learn how Salesforce successfully used the AWS Marketplace Private Offers program to procure an enterprise license and consolidate the expenditure onto their AWS bill.