United AI Studio

There’s a Review in progress. Changes made to the RUAI document will be saved, but will not be seen by the MLRB until a new Review is created.

Partitions allow for scaling of document counts as well as faster data ingestions by spanning your index over multiple units.

Replicas distribute workloads across the service. We guarantee 99.9% availability for read operations with 2 replicas, and for read and write operations with 3 or more replicas. Learn more about the SLA for Cognitive AI Search.

Add Azure AI Search

Approve the requested amount, or modify the amount to be approved.

You may approve this request as-is or make a modification to the approved quota.

Azure AI Vision gives your apps the ability to analyze images, read text, and detect faces with prebuilt image tagging, text extraction with optical character recognition (OCR), and responsible facial recognition. 

Azure AI Computer Vision

Document intelligence provides REST API access to an AI service that applies advanced machine learning to extract text, key-value pairs, tables, and structures from documents automatically and accurately. 

Azure AI Document Intelligence

REST API access to AI search Azure AI Search <link>(formerly known as "Azure Cognitive Search")</link> provides secure information retrieval at scale over user-owned content in traditional and generative AI search applications.

Azure AI Search

AI Speech enables to transcribe speech to text with high accuracy, produce natural-sounding text-to-speech voices, translate spoken audio, and use speaker recognition during conversations. 

Azure AI Speech

Azure Text Translation is a cloud-based REST API feature of the Translator service that uses neural machine translation technology to enable quick and accurate source-to-target text translation in real time across all <link>supported languages</link>.

Azure AI Translator

Unable to load Subscription Id. Please try again in a few minutes or reach out to support for assistance.

Unable to load completion form due to missing project data. Please try again in a few minutes or reach out to support for assistance.

ADF will be ready in a moment. Please wait.

If you have an Azure Application that you will be using with OpenAI, please provide the App ID for it below. To update the App ID after this step, you will need to make a support request.

Optionally provide the App ID for the Azure application you would like to use with OpenAI.

Configure Azure OpenAI

Name

Add Enterprise data to your project or Bring Your Own Data.

Are you sure you want to delete the Deployment?

Confirm Deletion

Deleting this endpoint will delete the unique Endpoint URI along with any other Endpoint information. There is no way to recover an endpoint after it has been deleted.

Are you sure you want to delete the Endpoint?

Active Deployment Count

Created on

Deployed Machine Type / SKU

Deployment ID

Description

Endpoint ID

Endpoint State

Environment

Unable to load deployments. Please try again in a few minutes or reach out to support for assistance.

Instance Count

REST endpoint

Scoring Script

Swagger URI

Type

Updated on

Listed below are all the model endpoints you have created in your project and their current status.

Unable to load endpoints. Please try again in a few minutes or reach out to support for assistance.

Error

You already have an active or pending request for {modelName}. <link>Click here</link> to view the request details.

Something went wrong

Something went wrong.

To enable Azure OpenAI for your project, please fill out the following configuration.

Learn more about the available OpenAI Models.

Select Model

Please enter the starting date on which you wish to use the model.

When do you need to start using the model?

Request OpenAI Model Access

Provide the client ID for the App Identity. You can get the client ID from the HCP console. If you don’t have one, please follow the steps to create an Application Identity.

Provide a nickname for the App Identity you are configuring.

Description must be 1024 characters or fewer.

Complete this form to request updates to the selected model. To request access to a different model

Current Details

Requested Updates

Request Updates to Model

To create a United AI Studio Project, you will need Contributor or Admin access to a <link>Healthcare Platform Resource Group</link>. To learn more, review <link2>How to create a project</link2> documentation.

Requested Options

Success

Add a Client ID from HCP Console App Identity.

Add App Identity

Authentication

Details

Request History

OpenAI - {modelName}

This subscription has met the limit for the number of Azure OpenAI instances that can be associated with it. If you need access to OpenAI for this project, please specify a subscription that has not exceeded this limit. (For example, a new subscription.) Please <link>contact support</link> if you need assistance.

Before adding OpenAI to your project, you must ensure that the <link>Responsible Use of AI form</link> has been completed. Once that has been filled, you will be able to add OpenAI to this Project.

Azure OpenAI Service provides REST API access to OpenAI’s powerful language models and Embeddings model series. These models can be easily adapted to your specific task, including but not limited to content generation, summarization, semantic search, and natural language-to-code translation. Users can access the service through REST APIs, Python SDK, or our web-based interface in the Azure OpenAI Studio.

Azure OpenAI

It looks like you have not completed all of the fields on this page. You can always access this form from Dashboard > Manage Project after your project has been successfully created.

You will need to complete all fields of the Responsible Use of AI form and submit to the Machine Learning Review Board (MLRB) before any ML models or analytics created within this project can be deployed.

Responsible Use of AI/ML

By moving on to the next step, you acknowledge that you will need to complete this Responsible Use of AI form in order to deploy your ML model or analytic.

Are you sure you want to discard this project?

You have filled in all the fields of the Responsible Use of AI Form. Great work! Your form is now marked COMPLETE which means all of the fields are filled and ready to be reviewed by the Machine Learning Review Board (MLRB). Feel free to continue modifying or adding to these fields as needed. Once you have published your models from your notebook, you can submit your project to the MLRB for review.

Project Created!

Data Products

Please ensure you have selected the correct Resource Group for your project as the selection cannot be changed in the future. For additional details, review <docsLink>How to create a Project in UAIS</docsLink>.

UAIS is now integrated with HCP. All new UAIS projects must first have an HCP Resource Group created. To create an HCP Resource Group, review the HCP documentation for <link>Account Manager</link> including:

A United AI Studio project encompasses the end-to-end ML lifecycle for a business use case, providing a consolidated view for data, models, deployed endpoints, monitoring, and cost transparency. A project can also be used to submit your externally developed model to the Machine Learning Review Board.

Please describe your project. This description will help the Machine Learning Review Board (MLRB) understand how your ML model will be used.

In partnership with Cross Optum, we’ve developed a machine learning algorithm that predicts healthcare costs for people with diabetes

A good project name contains both the problem you’re trying to solve and the client or executing team.

The PULSAR diabetes cost prediction group grants access to a UAIS environment used to build models that identify patients most likely to have high costs in the next year.

The SECURE group justification can be simple.

This group is necessary to grant collaborators access to UAIS resources in the cloud.

Project Details

Review & Submit

Responsible Use

Workspace Setup

Create a Project

Data composition

Data source

FHIR resources

Overview

Quality metrics

Sorry we can’t find that project. Head back to the homepage to explore all United AI Studio has to offer.

Project Navigation

Project Snapshot

Error fetching max available quota. Please contact support for assistance.

Please select a model to see available quota.

Corresponding Requests Per Minute (RPM) = {value}

Are you sure you want to remove this client ID?

Confirm Removal

No Services have been added to your project yet. Click the button below to add OpenAI to your project.

To add Azure OpenAI to this project, follow the steps outlined in this documentation page: <link>How to work with OpenAI</link>.

Complete the workspace creation process to begin adding services to your project.

OpenAI not enabled for this project.

To access your storage container, use the provided URL with popular tools like <azCopy>AzCopy</azCopy>, <storageExplorer>Azure Storage Explorer</storageExplorer>, or your code.

Unable to load review. Please try again in a few minutes or reach out to support for assistance.

There was an error updating your review. Please try again in a few minutes.

Databricks provides spark cluster management, allowing data science projects to efficiently process data using familiar languages like Python, R, and SQL.

Azure ML is the ideal tool for creating end-to-end machine learning pipelines in the cloud, including data processing, managing, and deploying ML-driven web services.

The United AI Studio (UAIS) Azure subscription displayed is connected to the AIDE ID of the Healthcare Platform (HCP) Resource Group selected when creating this Project.

Azure Databricks is currently available as an optional add-on to United AI Studio projects. More add-ons will be included here in the future.

The Azure ML platform is preselected for all United AI Studio projects that require a workspace.

Fixed Output

In-context Learning

Open-ended Output

Transfer Learning

Sorry, you do not have access to that page. Head back to the homepage to explore all United AI Studio has to offer.

This page is unauthorized.

Sorry we can’t find that page. Head back to the homepage to explore all United AI Studio has to offer.

This page doesn’t exist.

United AI Studio server is currently unable to handle this request. Try to refresh this page or contact us if the problem persists.

This page isn’t working.

OpenAI Model Requests

Helpful Links

Sorry, we couldn’t find anything that matched your search. Double check your search for typos or spelling errors, or try a different search term.

No results for “{searchTerm}” 

Models ({count})

Cloud Resources Cost Dashboard

The Cost Dashboard only includes the cloud resources consumption cost and is intended for early costing analysis and trends. It does not include reservations, shared cost, model deployment cost, OpenAI cost and other subscription charges included in actual Optum billing.

These dashboards helps the user understand the different costs associated with the azure resources consumption and helps the user to control and optimize costs.

Cloud Resources Usage Dashboard

To access your storage container, use the provided shared access signature (SAS) token or URL. Use your SAS with popular tools like <azCopy>AzCopy</azCopy> or <storageExplorer>Azure Storage Explorer</storageExplorer>.

This project was created for Machine Learning Review Board submission only and add-ons are not yet available.

Manage enterprise data transfers in your project.

Manage adhoc and recurring data transfers in your project.

Click the button above to add your first custom data set

Click the button above to add your first recurring data ingress.

Complete the workspace creation process to begin adding data to your project.

Your data submission is currently under review.

United AI Studio already has some valuable data products available from existing enterprise data warehouses made available to United AI Studio users in a way that enables managed updates and governance. This means you can be confident that your project’s data will not be exposed to unauthorized parties and won’t go out of date.

I plan to exclusively use my own data for my project.

Data in the form of files out of a UAIS Project

Recurring File Transfers RMDBS Data Export

I plan to exclusively use enterprise data for my project.

Once a model has been published for this project, it will appear here, along with its current status.

You have not published any models.

Listed below are the models you have published in your project and their current status, including MLRB review status.

Unable to load model version(s). Please try again in a few minutes or reach out to support for assistance.

Unable to load review(s). Please try again in a few minutes or reach out to support for assistance.

Any changes made will be saved, but they will not be incorporated into the active Review. Instead, the Review will need to be <link>withdrawn</link> and recreated to have those changes appear for the MLRB.

You have filled in all the fields of the Responsible Use of AI Form. Great work!Your form is now marked COMPLETE which means all of the fields are filled and ready to be reviewed by the Machine Learning Review Board (MLRB). Feel free to continue modifying or adding to these fields as needed.Once you have published your models from your notebook, you can submit your project to the MLRB for review.

Responsible Use of AI/ML form completed!

Responsible Use of AI Overview

Listed below are all the services enabled for your project.

There was an error updating your workspace options. Please try again in a few minutes.

The following workspaces have been requested and will be available soon. Workspaces may take up to 4 hours to initialize.

There was an error loading your Azure subscription options. Please try again in a few minutes.

Add cognitive AI services such as Azure OpenAI.

Sorry, we couldn’t find anything that matched your search. Double check your search for typos or spelling errors, or try different search options.

Search Projects

Projects

Identify the analytics team leader ultimately accountable for the development, testing, and ongoing monitoring of the AI/ML being considered.

Enter the name of the analytics team or department directly responsible for the development, testing, and ongoing monitoring of the ML being considered. 

Describe the anticipated benefits the AI/ML in this project will have on the impacted population groups, end users, and business segments who are implementing these solutions.

Provide a detailed description of the methodology used to assess the performance of the AI/ML in this project. Your response should specify what measures were identified for your use case and rationale for their selection. 

Provide the email address for the business contact who is implementing the AI/ML in this project.

Describe the dataset that was used to develop the AI/ML in this project. The total working dataset, testing dataset, and training dataset should be listed separately.

Describe which data sources were used to test, train and run the AI/ML in this project. Each data source should be listed with an easy to understand name without unexplained acronyms. Please indicate which data source was used for each AI/ML Solution in the project if there are multiple.

Select all decisions or actions taken as a result of the AI/ML in this project. Indicate “other” if your decision or action type is not listed. You will be prompted to provide a free-text response in the next question.

Select all end user types who will utilize the outputs of the AI/ML in this project. Indicate “other” if your end user type is not listed. You will be prompted to provide a free-text response in the next question.

If the AI/ML in this project includes AI/ML from an external vendor, supplier or other third parties, you will need to provide a downloadable version of this form to your external contacts to complete. A third party job aid is also available to share with vendors while they provide intake form responses. Contact the RUAI Program at ruai@uhg.com if you have any questions about what documentation may be shared externally.

Provide the name of the external vendor, supplier, or other third party who is involved with the AI/ML in this project. This should be the name of the company or organization rather than the names of your vendor contacts.

Describe your methodology to assess the outputs of the AI/ML in this project for fairness. As outlined in Phase III in the Guide for Responsible Development and Use of AI/ML, you must report all performance measures by Select Population Groups (SPGs) that apply to the use case. Your review request will be returned if fairness testing has not been completed.

Describe the results from your fairness assessment of the AI/ML in this project. Document what steps were taken to correct or address any potential bias findings. Clearly label your complete fairness assessment results and attach them to your submission. Your review request will be returned if your fairness assessment results are not clearly labeled and the methodology documented.

If you are seeking exemption from fairness testing, you will be asked to provide a written justification for this excemption request in the next question. The MLRB will review your exemption request to determine if fairness testing is not required for your use case. Note that it is uncommon for the MLRB to grant an exemption from fairness testing. Refer to Phase III in the Guide for Responsible Development and Use of AI/ML for the minimum fairness testing requirements.

If the AI/ML in this project includes a Generative AI or Large Language Model component, you will need to download a supplemental form with additional Generative AI questions. All required questions in the Generative AI Supplemental Intake form must be completed and attached to your submission before requesting MLRB review.

Describe how the data was evaluated for potential limitations and sources of bias, such as missingness, observational intensity, logistical limits, or low participation from different demographics. Document steps taken to address any bias found in the data and its impact on the AI/ML. Refer to Phase II in the Guide for Responsible Development and Use of AI/ML for the complete list of required steps and considerations while completing a data quality and data bias assessment for the AI/ML in this project.

In coordination with your business sponsors, identify whether the use of the AI/ML outputs from this project will affect individuals living within the United States and/or its territories.

In coordination with your business sponsors, identify whether the use of the AI/ML outputs from this project  will affect individuals living outside the United States and/or the US territories.

Select all applicable population groups impacted by the AI/ML in this project. Indicate “other” if your population group is not listed. You will be prompted to provide a free-text response in the next question.

Clearly document the use case the AI/ML in this project was designed to solve.

Provide an interpretation of your top features in each model to help explain why they are important. Attach the complete feature list and their importance metrics (for example, SHaP scores) to your submission)

Select all applicable business segments which are sponsoring or implementing the AI/ML in this project. Choices in this list align to the names of the External Segments found in the Business Hierarchy.

Describe how the AI/ML in this project will be monitored and assessed for performance, fairness, and business metric impact after deployment and on a regular cadence throughout the AI/ML lifecycle.

Select the frequency the AI/ML in this project will be regularly assessed for performance, fairness and business metric impact

Describe the internal governance group who provides ongoing oversight for the AI/ML in this project, including reviewing the performance and fairness assessment results to determine if any findings need to be mitigated.

Select the standard frequency the governance group will meet to provide ongoing oversight for the AI/ML in this project

Provide the name of each open-source application used in this project and a description of where the data was sourced.

Details of other impacted population types

Describe how the outputs of the AI/ML in this project are displayed to each end user type you listed above and attach an example of each output.

Signing in via Azure