AWS Machine Learning Blog

AWS AI Blog Month in Review: March 2017

We’ve just finished another month of AI solutions on the AWS AI Blog. Please take a look at our summaries below and learn, comment, and share. Thanks for reading!


Deploy Deep Learning Models on Amazon ECS
In this post, learn how to connect the workflow between the data scientists and DevOps. Using a number of AWS services, take the output of a model’s training and deploy it to perform predictions in real time with low latency and high availability. In particular, see the ease of deploying DL predict functions using Apache MXNet (a deep learning library), Amazon ECS, Amazon S3, and Amazon ECR, Amazon developer tools, and AWS CloudFormation.

Amazon at WMT: Improving Machine Translation with User Feedback
Since 2006, the important annual event Workshop on Machine Translation invites participants to submit machine translation systems for competitive ranking in a number of categories. This year Amazon, in collaboration with Germany’s Heidelberg University, is hosting a new competition for machine translation systems that adapt well to simulated customer feedback; in other words, systems that are able to correct their mistakes by learning from a stream of translation assessments. The results will be presented at this year’s WMT Conference in Copenhagen.

Build Your Own Text-to-Speech Applications with Amazon Polly
In this blog post, create a basic, serverless application that uses Amazon Polly to convert text to speech. The application has a simple user interface that accepts text in many different languages and then converts it to audio files which you can play from a web browser.

AI Tech Talk: How to Get the Most Out of Amazon Polly, a Text-to-Speech Service
Although there are many ways to optimize the speech generated by Amazon Polly‘s text-to-speech voices, new customers may find it challenging to quickly learn how to apply the most effective enhancements in each situation. The objective of this webinar is to educate customers about all of the ways in which they can modify the speech output, and to learn some insider tips to help them get the most out of the Polly service. This webinar will provide a comprehensive overview of the available tools and techniques available for modifying Polly speech output, including SSML tags, lexicons, and punctuation. The post has been updated with a link to the video archive for this presentation.

Updated AWS CloudFormation Deep Learning Template Adds New Features and Capabilities
We’ve updated the AWS CloudFormation Deep Learning template with exciting additional capabilities including automation to dynamically adjust the cluster to the maximum number of available worker instances when an instance fails to provision. This template also lets you choose between GPU and CPU instance types as well as adds support to run under either Ubuntu or Amazon Linux environments for your cluster. We’ve also added the ability to provision a new, or attach an existing Amazon EFS file system to your cluster to let you easily share code/data/logs and results.

Use Amazon Rekognition to Build an End-to-End Serverless Photo Recognition System
Amazon Rekognition is a service that makes it easy to add image analysis to your applications. With Rekognition, you can detect objects, scenes, and faces in images. You can also search and compare faces. Rekognition’s API lets you quickly add sophisticated deep learning-based visual search and image classification to your applications. In this post, we’ll focus on searching for objects and scenes in images.

AWS Collaborates With the National Science Foundation to Foster Innovation
Amazon Web Services and the National Science Foundation (NSF) are collaborating to foster innovation in big data research. Under the AWS Research Initiative (ARI) program, AWS and NSF will respectively support innovative research in the field of Big Data. With the advancements of techniques and technologies such as cloud-based Artificial Intelligence, Machine Learning, Big Data analytics and High-Performance Computing, ARI BIGDATA grants will help researchers maximize the value of their NSF grants to accelerate the pace of innovation.

Predicting Customer Churn with Amazon Machine Learning
Losing customers is costly for any business. Identifying unhappy customers early on gives you a chance to offer them incentives to stay. This post describes using machine learning (ML) for the automated identification of unhappy customers, also known as customer churn prediction. ML models rarely give perfect predictions though, so the post is also about how to incorporate the relative costs of prediction mistakes when determining the financial outcome of using ML.

Deep Learning AMI release v2.0 now Available for Amazon Linux
You can now use upgraded versions of Apache MXNet, TensorFlow, CNTK, and Caffe, on the AWS Deep Learning AMI v2.0 for Amazon Linux, including Keras, available in the AWS Marketplace. The Deep Learning AMI v2.0 for Amazon Linux is designed to continue to provide a stable, secure, and high performance execution environment for deep learning applications running on Amazon EC2. The latest MXNet release (v0.9.3) included with this AMI v2.0 adds several enhancements including a faster new image processing API that enables parallel processing, improved multi GPU performance and support for new operators.

AI Tech Talk: Introducing Amazon Lex – A Service for Building Voice or Text Chatbots
Amazon Lex is a service for building conversational interfaces into any application using voice and text. Lex provides the advanced deep learning functionalities of automatic speech recognition (ASR) for converting speech to text, and natural language understanding (NLU) to recognize the intent of the text, to enable you to build applications with highly engaging user experiences and lifelike conversational interactions. The post has been updated with a link to the video archive for this presentation.

Leave a comment below to let us know what topics you’d like to see next on the AWS AI Blog.