Skip to content

Latest commit

 

History

History
64 lines (44 loc) · 5.43 KB

File metadata and controls

64 lines (44 loc) · 5.43 KB
title Data science roles and permissions
description Understand the permissions required for machine learning models and experiments in [!INCLUDE [product-name](../includes/product-name.md)] and how to assign these permissions to users.
ms.author scottpolly
author s-polly
ms.reviewer ruxu
reviewer ruixinxu
ms.topic concept-article
ms.custom
ms.date 02/26/2026

Data science roles and permissions

This article describes machine learning model and experiment permissions in [!INCLUDE product-name] and how users acquire these permissions.

Note

After you create a workspace in [!INCLUDE product-name], or if you have an admin role in a workspace, you can give others access to it by assigning them a different role. To learn more about workspaces and how to grant users access to your workspace, see Workspaces and Giving users access to workspaces.

Permissions for machine learning experiments

The following table describes the levels of permission that control access to machine learning experiments in [!INCLUDE product-name].

Permission Description
Read Allows user to read machine learning experiments.
Allows user to view runs within machine learning experiments.
Allows user to view run metrics and parameters.
Allows user to view and download run files.
Write Allows user to create machine learning experiments.
Allows user to modify or delete machine learning experiments.
Allows user to add runs to machine learning experiments.
Allows user to save an experiment run as a model.

Permissions for machine learning models

The following table describes the levels of permission that control access to machine learning models in [!INCLUDE product-name].

Permission Description
Read Allows user to read machine learning models.
Allows user to view versions within machine learning models.
Allows user to view model version metrics and parameters.
Allows user to view and download model version files.
Write Allows user to create machine learning models.
Allows user to modify or delete machine learning models.
Allows user to add model versions to machine learning models.

Permissions acquired by workspace role

A user's role in a workspace implicitly grants them permissions on the datasets in the workspace, as described in the following table.

Admin Member Contributor Viewer
Read :::image type="icon" source="./media/data-science-overview/yes.png" border="false"::: :::image type="icon" source="./media/data-science-overview/yes.png" border="false"::: :::image type="icon" source="./media/data-science-overview/yes.png" border="false"::: :::image type="icon" source="./media/data-science-overview/yes.png" border="false":::
Write :::image type="icon" source="./media/data-science-overview/yes.png" border="false"::: :::image type="icon" source="./media/data-science-overview/yes.png" border="false"::: :::image type="icon" source="./media/data-science-overview/yes.png" border="false"::: :::image type="icon" source="./media/data-science-overview/no.png" border="false":::
Execute :::image type="icon" source="./media/data-science-overview/yes.png" border="false"::: :::image type="icon" source="./media/data-science-overview/yes.png" border="false"::: :::image type="icon" source="./media/data-science-overview/yes.png" border="false"::: :::image type="icon" source="./media/data-science-overview/no.png" border="false":::
View execution output :::image type="icon" source="./media/data-science-overview/yes.png" border="false"::: :::image type="icon" source="./media/data-science-overview/yes.png" border="false"::: :::image type="icon" source="./media/data-science-overview/yes.png" border="false"::: :::image type="icon" source="./media/data-science-overview/yes.png" border="false":::

Note

You can assign roles to individuals or to security groups, Microsoft 365 groups, and distribution lists. For more information about workspace roles in [!INCLUDE product-name], see Roles in workspaces.

Item-level sharing

In addition to workspace roles, you can share individual machine learning models and experiments with specific users or groups. Sharing grants the recipient Read permission by default, with the option to grant additional permissions such as Edit, Share, or Execute. Use item-level sharing when you want to grant access to a specific item without assigning a workspace role.

For more information, see [Share items in [!INCLUDE product-name]](../fundamentals/share-items.md).

Related content