Amazon Web Services (AWS) is the world's leading cloud provider, and this week during its AWS Re:Invent conference in Las Vegas, the company unveiled new offerings that may not only help it stay on top, but also push cloud-based artificial intelligence (AI) workloads further into the IT mainstream.
AWS unveiled no less than five new machine learning services at Re:Invent, plus what Amazon claims is the first developer-focused, deep learning-enabled video camera. Dubbed AWS DeepLens, the offering combines hardware, on-board AI processing and cloud connectivity, allowing developers to train their world-sensing AI models out in the wild.
On the hardware front, AWS DeepLens includes a Wi-Fi-connected 4 megapixel, 1080P camera, an Intel Atom processor, 8 GB of memory and a 2D microphone array. The hardware runs Ubuntu 16.04 ships with Greengrass Core, software that enables IoT interoperability. It also includes an optimized version of the MXNet deep learning framework and Intel's clDNN library, which helps accelerate AI processing on Atom chips.
For AI developers who want to hit the ground running, Amazon is also offering pre-trained face and object recognition models. These components form a system that can turn the world at large into an AI testbed, according to Jeff Barr, chief evangelist at AWS.
"All of this hardware, software, and data come together to make the AWS DeepLens a prime example of an edge device," blogged Barr. "With eyes, ears, and a fairly powerful brain that are all located out in the field and close to the action, it can run incoming video and audio through on-board deep learning models quickly and with low latency, making use of the cloud for more compute-intensive higher-level processing. For example, you can do face detection on the DeepLens and then let Amazon Rekognition take care of the face recognition."
AWS DeepLens hardware costs $249.00 and is slated to ship on April 14, 2018. Amazon is currently accepting preorders here.
Amazon Rekognition Video is a new deep-learning based video recognition system that allows developers to add object, scene and facial recognition, among several other parameters, to their video-based applications. Another new AI offering from AWS is SageMaker, a toolset that accelerates the machine learning model building, training and hosting process, allowing businesses to quickly add AI capabilities to their production applications. SageMaker is made up of hosted Jupyter notebook IDEs (integrated development environments), a distributed system for model building, training and validation, and finally, a hosting component with HTTP endpoints.
The AWS AI suite has also developed a knack for languages.
Amazon Translate is a new, real-time neural machine language translation service that can be linked with other services like AWS Elasticsearch for multi-language search and Amazon Lex for translation chatbots. The company's new continuously-trained natural language processing service, called Amazon Comprehend, analyzes text, identifies entities, gauges sentiment and extracts topics from text-based information.
Finally, Amazon Transcribe is an automatic speech recognition service allows developers to add speech-to-text capabilities to their applications. For now, the preview service can transcribe speech in English and Spanish.
Pedro Hernandez is a contributing editor at Datamation. Follow him on Twitter @ecoINSITE.