Azure at Microsoft Ignite 2022: Key updates for the Microsoft cloud community

October 12 2022

Microsoft Ignite 2022 is here. As in years past, Microsoft has released a broad slate of Azure updates and announcements in conjunction with the event, signaling where its cloud roadmap is heading next. Let’s take a look at some of these major updates—and what they mean for Azure in the months to come.

Cognitive Services

Cognitive Services received a wide variety of updates. Azure OpenAI Service is out in limited availability preview, with access to Dall E-2, a custom image generator app. Using Dall E, customers can generate code and content as needed. Cognitive Service for Speech added Neural Text to Speech to scale up voice assistants, synthetic voice training for Custom Neural Voice, and new audio processing for prebuilt voices. Along similar lines, Cognitive Service for Language added natural language understanding and natural language generation options for document and conversation digitization scenarios.  

The Computer Vision portfolio added two new services. Image Analysis 4.0 has an updated model intended to extract more visual image data to boost digital asset management, whereas Spatial Analysis on the Edge ingests video streams for analysis. Microsoft broadened Form Recognizer to encompass 275 languages and include a prebuilt AI model for contract data extraction.

AI & ML

Microsoft signaled a growing commitment to cross-team ML projects with the preview of Machine Learning registries. With this option, ML developers can discover and share ML models or pipelines across multiple organization workspaces, and thereby track model and data lineage. For PyTorch modeling, Microsoft created previews of Container for PyTorch and Data Science Virtual Machines for PyTorch.

Building off announcements earlier this year about its Responsible AI initiatives, Microsoft has introduced a Responsible AI scorecard that generates PDF scorecards for no code and low code customers to report on the compliance of their models. The team also announced Python SDK v2, automated machine learning features like natural language processing, and Apache Spark pools integration.

Microsoft’s Project Bonsai low-code industrial AI platform, currently in preview, added several preview features. For instance, it now supports training with OfflineRL data, model-based training acceleration, and action masking, restricting actions based on the current system state. Assessment 2.0 is a new tool to visualize policy trajectories and track low-performance episodes. Going forward, Tata Consultancy Services is a partner able to support Project Bonsai.

Databases & Analytics

Azure database and analytics capabilities always garner a lot of attention at Microsoft events, and Ignite 2022 is no exception. Microsoft highlighted partnerships for Intelligent Data Platform: MongoDB and YugabyteDB New SQL and NoSQL support; Informatica, Confluent, dbt Labs, Fivetran, Qlik and Striim low code/no code integrations; and Profisee, CluedIn, Delphix and OneTrust multicloud data estate support.

Using the Hyperscale (Citus) engine, Microsoft extended distributed PostgreSQL support to Cosmos DB. Developers typically use Azure Database for Cosmos DB to set up relational and non-relational databases. With the changes, developers can store documents that are more than 16 MB in Cosmos DB for MongoDB.

Microsoft added support for Oracle Database assessments in Data Studio, issued Synapse Analytics pipeline templates for Microsoft 365 data, and set Data Explorer to run real-time analytics for Event Grid, Amazon S3, and OpenTelemetry metrics. Simultaneously, it launched Purview Data Governance as a way to conduct root cause analysis, metamodeling, and ML-based classifications.

Hybrid, Edge & Multicloud

About MSCN Reporter

More about MSCN Reporter