Microsoft Syntex adds taxonomy and image tagging, OCR, content query, annotations, and more

by Contributed | Oct 11, 2023 | Technology

This article is contributed. See the original author and article here.

Welcome to the fall! This month’s Microsoft Syntex update is gearing up to be a great one in the world of content processing. We have updates on Syntex taxonomy tagging and image tagging; a set of Syntex capabilities coming to preview for pay-as-you-go users; the general availability of the Syntex optical character recognition (OCR) is expanding to include PDF and TIFF support; and lastly, both Syntex OCR and Syntex structured document processing are moving to general availability.

Syntex taxonomy tagging and Syntex image tagging in general availability

In our previous blog post, we shared that Syntex taxonomy tagging and image tagging were rolling into general availability. We’re happy to share that both services have now completed rollout and are generally available to all Syntex pay-as-you-go users.

As a refresher, Syntex Taxonomy Tagging uses AI to help you label and organize documents by automatically tagging them with descriptive keywords, based on your taxonomy defined in SharePoint. By applying a taxonomy column and enabling taxonomy tagging, the document is automatically tagged with keywords from your term store to help with searching, sorting, filtering and more. This reduces manual work, and makes it faster and more efficient to categorize, find, and manage files in your document libraries. Overview of taxonomy tagging in Microsoft Syntex – Microsoft Syntex | Microsoft Learn

Taxonomy Tagging – the location column auto-populates based on your term store in this example

Syntex Image Tagging is now also generally available. Image Tagging is an AI-powered service that helps you label and organize images by automatically tagging them with descriptive keywords. These tags are stored as metadata to optimize searching, sorting, filtering, and managing your images. With this Syntex service, it’s much faster to categorize and search for specific images that you need. Overview of enhanced image tagging in Microsoft Syntex – Microsoft Syntex | Microsoft Learn

Image Tagging – images are auto-tagged with descriptive keywords

New features in preview for Syntex pay-as-you-go users

We’re excited to share that, for a limited time, Syntex pay-as-you-go users now get to use all the Syntex features previously only available to customers with the SharePoint Syntex seat license. If you’re not yet a Syntex customer, now is a particularly great time to give it a go. These services will be available as a preview through June 30, 2024.

1. Content query – an advanced, powerful search with custom metadata in a form-based interface
2. Universal annotation – add ink and highlights to additional file types like PDF & TIFF supported by our file viewer
3. Accelerators – preconfigured templates that leverage Syntex capabilities in an end-to-end solution for common scenarios like contract management and accounts payable
4. Taxonomy services – admin reporting on term set usage, easy import from SKOS-formatted taxonomies, and the ability to push a content type to a hub
5. Content processing rules – lightweight automation for common operations such as moving or copying a file, and setting a content type from the file name or path in SharePoint
6. PDF merge/extract – combine two or more PDF files into a new PDF file, or extract pages from one PDF into a new one

Site accelerator – preconfigured site templates for Accounts Payable

For Syntex pay-as-you-go customers, these capabilities will be readily available to you without taking any action or set up and without any additional charge. Microsoft Syntex features limited time license – Microsoft Syntex | Microsoft Learn

Syntex OCR and structured document processing will be generally available this month

Lastly, Syntex optical character recognition (OCR), which was previously in public preview, will be generally available this month! In images containing text – such as screenshots, scanned documents, or photographs – Syntex OCR automatically extracts the printed or handwritten text and makes it discoverable, searchable, and indexable.

It can be used for image-only files, now including PDF and TIFF as mentioned in the introduction, in OneDrive, SharePoint, Exchange, Windows devices and Teams messages. Searching for images is improved thanks to OCR, and IT admins can better secure images across OneDrive, SharePoint, Exchange, Teams and Windows devices with data loss prevention (DLP) policies.

Optical Character Recognition (OCR) auto-extracts text from images

You will also be able to use both the Syntex structured and freeform document processing features later this month when they become generally available as a new pay-as-you-go meter called “Structured document processing”. Unlike in the past, you will be able to use these services with your Azure subscription – no per-user license required, no AI Builder credits needed (but if you want to use AI Builder credits, we will still support that as well). Microsoft 365 roadmap ID 167309. Get started here.

Stay connected

And there you have it, lots of updates on Syntex services that will help your organization manage your content and improve content discovery, with less redundancy and greater efficiency. To get the latest on Syntex, join our mailing list for updates, and register for the upcoming October 18th Syntex Community Call.

Be sure to also connect with us at Microsoft Ignite, November 14-17, 2023, in Seattle or virtually!

Realize Lakehouse using best of breed of Open source using HDInsight

by Contributed | Oct 10, 2023 | Technology

This article is contributed. See the original author and article here.

Author: Reems Thomas Kottackal, Product Manager

HDInsight on AKS is a modern, reliable, secure, and fully managed Platform as a Service (PaaS) that runs on Azure Kubernetes Service (AKS). HDInsight on AKS allows an enterprise to deploy popular open-source analytics workloads like Apache Spark, Apache Flink, and Trino without the overhead of managing and monitoring containers.

You can build end-to-end, petabyte-scale Big Data applications spanning event storage using HDInsight Kafka, streaming through Apache Flink, data engineering and machine learning using Apache Spark, and Trino‘s powerful query engine. In combination with Azure analytics services like Azure data factory, Azure event hubs, Power BI, Azure Data Lake Storage.

HDInsight on AKS can connect seamlessly with HDInsight. You can reap the benefits of using needed cluster types in a hybrid model. Interoperate with cluster types of HDInsight using the same storage and meta store across both the offerings.

The following diagram depicts an example of end-end analytics landscape realized through HDInsight workloads.

We are super excited to get you started, lets get to how?

• Signup today – https://aka.ms/starthdionaks


• Read our documentation – https://aka.ms/hdionaks-docs


• Join our community, share an idea or share your success story – https://aka.ms/hdionakscommunity


• Have a question on how to migrate or want to discuss a use case – https://aka.ms/askhdinsight 

Easily Bring your Machine Learning Models into Production with the AzureML Inference Server

by Contributed | Oct 9, 2023 | Technology

This article is contributed. See the original author and article here.

Taking your machine learning (ML) models from local development into production can be challenging and time consuming. It requires creating a HTTP layer above your model to process incoming requests, integrate with logging services, and handle errors safely. What’s more, the code required for pre- and post-processing, model loading, and model inference vary across models and must integrate smoothly with the HTTP layer. 

Today, we are excited to announce the General Availability (GA) and open sourcing of the Azure Machine Learning Inference Server. This easy-to-use python package provides an extensible HTTP layer that enables you to rapidly prepare your ML models for production scenarios at scale. The package takes care of request processing, logging, and error handling. It also provides a score script interface that allows for custom, user-defined, pre- and post-processing, model loading, and inference code for any model. 

Summary of the AzureML Inference HTTP Server 

The Azure Machine Learning Inference Server is a Python package that exposes your ML model as a HTTP endpoint. The package contains a Flask-based server run via Gunicorn and is designed to handle production scale requests. It is currently the default server used in the Azure Machine Learning prebuilt Docker images for inference. And, while it is built for production, it is also designed to support rapid local development. 

Figure 1: How the Azure Machine Learning Inference Server Handles Incoming Requests

Score Script 

The score script (sometimes referred to as the “scoring script” or “user code”) is how you can provide your model to the server. It consists of two parts, an init() function, executed on server startup, and a run() function, executed when the server receives a request to the “/score” route.  

On server startup… 

The init() function is designed to hold the code for loading the model from the filesystem. It is only run once. 

On request to “/score” route… 

The run() function is designed to hold the code to handle inference requests. The code written here can be simple: passing raw JSON input to the model loaded in the init() function and returning the output. Or, it can be complex: running several pre-processing functions defined across multiple files, delegating inference to a GPU, and running content moderation on the model output before returning results to the user. 

The score script is designed for maximum extensibility. Any code can be placed into init() or run() and it will be run when those functions are called as described above. 

An example score script can be found here – azureml-examples – Github. 

Key AzureML Inference Server Scenarios 

Local Score Script Development and Debugging

Developing a complex score script may require iterative debugging and often it’s not feasible to redeploy an online endpoint several times to debug potential issues. The AzureML Inference Server allows you to run a score script locally to test both model loading and inference request handling. It easily integrates with the VS Code debugger and allows you to step through potentially complex processing or inference steps. 

Note: To test your docker image in addition to your score script please refer to the AzureML Local Endpoints documentation. 

Validation gates for CI/CD Pipeline 

The Azure Machine Learning Inference Server can also be used to create validation gates in a continuous integration and deployment (CICD) pipeline. For example, you can start the server with a candidate score script and run a test suite against this local instance directly in the pipeline, enabling a safe, efficient, and automatable deployment process.

Production Deployments 

The Azure Machine Learning Inference Server is designed to support production-scale inference. Once local testing is complete, you can feel confident using the score script you developed alongside the Azure Machine Learning prebuilt inference images to deploy your model as an AzureML Managed Online Endpoint. 

Safely bring your models into production using the Azure Machine Learning Inference Server and AzureML Managed Inference by referencing the resources below. 

Learn More 

• Azure Machine Learning (AzureML) Inference Server Documentation: Azure Machine Learning inference HTTP server – Azure Machine Learning | Microsoft Learn 

• Azure Machine Learning (AzureML) Inference Server Open-Source GitHub repository: microsoft/azureml-inference-server: The AzureML Inference Server is a python package that allows user to easily expose machine learning models as HTTP Endpoints. The server is included by default in AzureML’s pre-built docker images for inference. (github.com) 


• Azure Machine Learning (AzureML) Inference Server PyPI package: azureml-inference-server-http · PyPI