AI + Machine Learning

Three new features now available in Public Preview enable you to utilize vision-based models in the model catalog, enjoy seamless integration with LLMOps through the Prompt Flow SDK/CLI and VS…

Two new features now available in PuPr enable you to work with improved data size and model support and generate a default Responsible AI dashboard for the best AutoML model.

New features now available in Public Preview enable you to create AI workflows that connect to various language models and data sources and use prompt flow and Azure OpenAI models…

A new feature now available in GA enables you to create compute clusters in locations that are different from the location of the workspace.

Azure Video indexer will be adding Advanced Audio and Video SKUs for Advanced presets and a 40% price reduction on the Basic Audio Analysis, Audio Analysis and Video Analysis SKUs.

New feature now available in Public Preview enables you to view job statuses without unnecessary compute cost accrual.

Azure OpenAI Service is now available in the UK South region. UK South supports gpt-35-turbo at launch.

Azure AI Content Safety is a new Azure AI service that identifies harmful online content, categorizes it and assigns severity scores to support human content moderator workloads.

New features now available in PuPr include the ability to evaluate text/image data fairness and enjoy one platform to build, tune, evaluate, deploy, and test AI workflow with LLM. 

New features now available in GA include the ability to enable MLOPs across dev, test, and prod environments to promote assets across multiple workspaces.

We recently made the Speech-to-text REST API version 3.1 generally available. With this release, we'll retire the preview version v3.1-preview.1 on May 12, 2023.

Continual Language ID makes transcription and translation easier by identifying up to 10 languages in a single piece of audio continuously.

New features now available in GA include the ability to customize your compute instance and configure a compute instance to automatically stop if it is inactive.

New features now available in GA include the ability to visualize timeseries models, and create a Compute Instance on behalf of another user.

New feature now available in Public Preview includes the ability to receive troubleshooting documentation on failed environment builds and shorten the training phase of large-scale distributed PyTorch models. 

Azure Machine Learning is now Generally Available in the UK West region.

Build conversational AI experiences with Azure Communication Services. Use virtual chat agents to manage growing customer service needs and provide in-chat intelligence with the help of smart assistants.

New features now available in Public Preview include the ability to build an end-to-end training pipeline with no-code, recover a deleted workspace, and identify root causes of pipeline failure.

This insight provides information on key items worn by individuals within a video and the timestamp in which the clothing appears.

With an Azure free account, you can explore with free amounts of 55+ always free services.

New public preview features include reading Delta Lakes in fewer steps, debugging and monitoring training jobs, and performing data wrangling.

New GA features include the ability to automate auto-shutdown/auto-start schedules, configure and customize a compute instance, seamlessly build NLP/vision models, and assess AI systems.

Multivariate Anomaly Detection will allow you to evaluate multiple signals and the correlations between them to find sudden changes in data patterns.

Features include functionality to promote pipelines and models across workspaces, perform data wrangling at scale, shorten training times, and lower set-up costs with Azure Container for PyTorch.

Features include controlling/customizing a model’s training code and using Python functions to develop tasks.

New features include the ability to establish event-driven notifications and the capability to label data in text documents using text named-identity recognition.

Includes the ability to control access to sensitive data, the capability to contrast differences to assess their performance, and the functionality to stop idle compute instances automatically.

Hierarchical forecasting, now generally available, offers you the capability to produce consistent forecasts for all levels of your data.

New features now available in public preview include AutoML Code Generation enhancements and AutoML in Pipelines.

The Cognitive Services team has updated our approach to facial recognition including adding a new Limited Access policy, removing AI classifiers of sensitive attributes and bolstering our investments in fairness…

New features include MLflow enhancements and train and deploy models in Azure hybrid and multi-cloud.

Conversational language understanding includes multilingual transformer-based models as the underlying model architecture and results in significant accuracy improvements over its predecessor, LUIS.

Enhancements in preview include Power Virtual Agents to enable sophisticated conversational flows; Azure Form Recognizer helps to streamline patient check-in and verification; Azure Metrics Advisor has simplified service customization.

Azure Cognitive Services is announcing the following preview updates at Build: Azure OpenAI Service and summarization capabilities in Azure Cognitive Service for Language.

Cognitive Service for Language now includes custom named entity recognition and custom text classification.

Deploy your machine learning models in a turnkey manner across powerful GPU and GPU machines. CLI v2 now also generally available.

Azure Machine Learning now offers a Responsible AI dashboard and scorecard as well as additional enhancements in preview.

Computer Vision Read API for Optical Character Recognition (OCR) announced the general availability of the new model with support for 164 languages. It’s also available as a Docker container.

Label data with text classification using machine learning assist.

New features include multi-class, multi-label, and named entity recognition, search for Azure Machine Learning assets across multiple workspaces, and label with text named entity recognition.

Compartmentalize access to data with identity passthrough and configure email notifications from Azure Learning.

Start using Ray on Azure Machine Learning Library before 30 June 2022.

Streamline your Cognitive Search operations with debug sessions for building, testing, and modifying skillsets.

Azure Form Recognizer now supports extraction from custom documents, W-2 forms, access to the read API, additional human languages and enhancements.

New features now available in public preview include automatic time series ID detection for automation, automatedML model’s training code generation, and move Azure Machine Learning workspaces between subscriptions.

Start using Anomaly Detector’s multivariate detection synchronously for quicker response, easier operations, and more informative result interpretation to help you analyze root causes.

New features now available in General Availability include add multiple private endpoints for workspaces, and support for storage account IP firewall.

Semantic Search, a feature of Azure Cognitive Search, is now available for direct configuration in the Azure Portal.

A new feature now available in general availability includes Manage Azure Machine Learning resource using Hashicorp Terraform

The new feature in general availability for Azure Machine Learning includes deploying with pre-built Inferencing docker images.