In-Depth Analysis Alibaba Cloud (also known as “Aliyun”) introduced its Ali-Perseus unified and distributed communication framework for deep learning using TensorFlow, Caffe, PyTorch, and MXNet. It compared Ali-Perseus cloud scaling performance against the LF AI Foundation’s open source Horovod distributed training framework for TensorFlow, Keras, PyTorch, and MXNet. It is likely that Alibaba is already using Ali-Perseus to scale it’s own production deployments of AI training and inference tasks. Amazon Web Services (AWS) announced a new feature, StyleSnap, for Amazon Fashion. StyleSnap is an AI-powered feature using computer vision and deep learning designed to make shopping easier. Amazon users can now take a photo or screenshot of a look they like, select “StyleSnap” in the Amazon app and then upload the photo. From there, StyleSnap will recommend similar items available in Amazon’s marketplace that match the look of the photo. The feature competes with a similar image recognition software from Alibaba that also allows users to input images to find products and fulfill search requests. Alibaba’s software was heavily discussed as a differentiator for the company in DoubleHorn’s March 2019 webinar ‘Same’ Does Not Mean Equal. Finally, AWS announced Gluon Time Series (GluonTS), an Apache MXNet-based toolkit for time series analysis using the Gluon API, is available as open source software on Github. Google Cloud experienced a major outage on Sunday June 2. Google Cloud reported “the root cause…was a configuration change that was intended for a small number of servers in a single region…the configuration was incorrectly applied to a larger number of servers across several neighboring regions, and it caused those regions to stop using more than half of their available network capacity.” Google Cloud’s acquisition of Looker is the first major acquisition under Google Cloud Chief Executive Thomas Kurian. Looker is designed to enable analysts define calculations to visualize data trends without writing complicated scripts. Google also announced it is launching a capacity reservation system for Compute Engine that allows users to reserve resources in a specific zone for later use. This system guarantees access to those resources when needed such as for disaster recovery. Snowflake expanded its data warehouse product offering to integrate with Google Cloud Platform. Snowflake promises seamless and secure data integration for organizations and across platforms through its cloud-built data warehouse Microsoft Azure’s big announcement last week was its new partnership with Oracle Cloud. The companies reached an agreement to create high-speed links between their data centers, making their two cloud computing services work together. This partnership is the first time we are seeing two major cloud providers collaborate in such a way. Oracle Cloud and Azure said the high-speed link that will run between their data centers will begin with facilities located in the Eastern US and then spread to other regions.They will also work together to allow joint users to log into services from either company with a single username, as well as get tech support from either company. Azure’s new Verifier for Solidity tool pairs with Azure’s blockchain suite and, according to the company, streamlines the code auditing process with automated security checks. Microsoft’s new tool seeks to prevent small logical error occurrences and hacks from happening. Azure also announced new capabilities and accuracy for time-series forecasting in Azure Machine Learning service. Some of the new capabilities include expanding forecast function and rolling-origin cross-validation, among others. Finally, Azure announced mobility service for Azure maps and SDK updates. Azure’s mobility service will start by “powering public transit routing, enabling organizations to add public transportation information and routing capabilities into their mobility, IoT, logistics, asset tracking, smart cities, and similar solutions.” | 
