Update: Microsoft will be moving away from UserVoice sites on a product-by-product basis throughout the 2021 calendar year. We will leverage 1st party solutions for customer feedback. Learn more here.

Machine Learning

Welcome! The Azure Machine Learning team invites you to share and vote for features to help you build, manage, and deploy custom machine learning models.

Have a technical question, or want to learn more? Please visit our documentation, MSDN forum or StackOverflow.


  1. AutoML (and designer) should post best hyperparameters in run results

    AutoML are useful now in finding the best model, but it should also report the hyperparameters for that best model (since spends time finding those).

    Maybe it (and the designer activity called tune hyperparameters) could return the results of the #leaves, #branches, etc) in the experiment run results.

    1 vote
    Vote

    We're glad you're here

    Please sign in to leave feedback

    Signed in as (Sign out)
    You have left! (?) (thinking…)
    0 comments  ·  Flag idea as inappropriate…  ·  Admin →
  2. interface cluster (gui) should allow creation of ACI clusters

    Right now the interface cluster gui only seems to allow AKS clusters to be created.

    1 vote
    Vote

    We're glad you're here

    Please sign in to leave feedback

    Signed in as (Sign out)
    You have left! (?) (thinking…)
    0 comments  ·  Flag idea as inappropriate…  ·  Admin →
  3. rename clusters

    I don't see a way to rename a compute cluster

    1 vote
    Vote

    We're glad you're here

    Please sign in to leave feedback

    Signed in as (Sign out)
    You have left! (?) (thinking…)
    0 comments  ·  Flag idea as inappropriate…  ·  Admin →
  4. Azure Auto ML Clustering For Mixed Data Types - Categorical & Numeric

    Azure Auto ML presently does not support clustering for mixed data types . Open source projects at the moment (like one in R using Gower Distance) presently provide it and this is much more practical than just having clustering based on Numeric(e.g. Euclidean Distance) or just Categorical(e.g. Jaccard Distance).

    1 vote
    Vote

    We're glad you're here

    Please sign in to leave feedback

    Signed in as (Sign out)
    You have left! (?) (thinking…)
    0 comments  ·  Flag idea as inappropriate…  ·  Admin →
  5. Azure Machine Learning A/B Testing Feedback Collection Framework

    Azure Machine Learning should support an established framework to collect model performance feedback based on real time production execution rather than just the initial statistical evaluation provided by Data Scientist. This also needs to integrate with AML Experiments/AML Model Registry so that in cases where models are being trained frequently (due to less data for example) , the system can automatically identify that A/B testing is not good and the previously successfully running model can be automatically be reverted to. The reversal would also have to be coordinated with Data Scientists possible in form of approval notifications

    1 vote
    Vote

    We're glad you're here

    Please sign in to leave feedback

    Signed in as (Sign out)
    You have left! (?) (thinking…)
    0 comments  ·  Flag idea as inappropriate…  ·  Admin →
  6. Azure Machine Learning Designer In AML Studio Model Evaluation, Curves Not Adjusting With Threshold Change

    AUC curves do not adjust in Evaluate Model result visualization along with the threshold adjustment although the confusion matrix does

    1 vote
    Vote

    We're glad you're here

    Please sign in to leave feedback

    Signed in as (Sign out)
    You have left! (?) (thinking…)
    1 comment  ·  Flag idea as inappropriate…  ·  Admin →
  7. Recovery of Azure ML Workspace

    Hi Team,

    Can you implement in Azure to recover Machine Learning Workspaces which gets deleted accidently by developers?

    Since once they delete Workspaces, Storage account associated to those workspaces also gets deleted and they lose their data.

    Thanks,
    Bhagat

    1 vote
    Vote

    We're glad you're here

    Please sign in to leave feedback

    Signed in as (Sign out)
    You have left! (?) (thinking…)
    0 comments  ·  Flag idea as inappropriate…  ·  Admin →
  8. 1 vote
    Vote

    We're glad you're here

    Please sign in to leave feedback

    Signed in as (Sign out)
    You have left! (?) (thinking…)
    0 comments  ·  Flag idea as inappropriate…  ·  Admin →
  9. Add options to setup Pod's memory and CPU on model deployment through Designer

    After submitting a real-time inference pipeline, when I click on deploy I have not any option to choose memory and CPU. This cause problems on heavy models, leading to an endpoint with a no ending transitioning state. This issue can be outlined only through a manual increasing of the pods memory.

    1 vote
    Vote

    We're glad you're here

    Please sign in to leave feedback

    Signed in as (Sign out)
    You have left! (?) (thinking…)
    1 comment  ·  Flag idea as inappropriate…  ·  Admin →
  10. HyperDrive: Allow reusing child runs even when parameter space changes

    I had a hyperdrive run going, and it became clear that one of my parameter's lower bound was too high (i.e. the best runs were all at that lower bound), so I killed the HD run and tried to warmstart a new one with just that run, but it rejected that because the space had changed.

    It would be nice if I could reuse the child runs from a prior HD run and just discard the ones that fall outside of the new parameter space.

    1 vote
    Vote

    We're glad you're here

    Please sign in to leave feedback

    Signed in as (Sign out)
    You have left! (?) (thinking…)
    0 comments  ·  Flag idea as inappropriate…  ·  Admin →
  11. Custom Module in Designer

    Include a custom module in Designer where people can call an existing model from the registry and build around it, or call web service of a model outside of Designer and build around it, or call a cognitive service and build around it or call an ONNX model and build around it.

    1 vote
    Vote

    We're glad you're here

    Please sign in to leave feedback

    Signed in as (Sign out)
    You have left! (?) (thinking…)
    0 comments  ·  Flag idea as inappropriate…  ·  Admin →
  12. Display and Edit Python Code by Module

    It would be terrific to view the python code being generated for a module and make tweaks to it in a code editor.

    1 vote
    Vote

    We're glad you're here

    Please sign in to leave feedback

    Signed in as (Sign out)
    You have left! (?) (thinking…)
    0 comments  ·  Flag idea as inappropriate…  ·  Admin →
  13. Numbered Diagrams

    Hi,

    It is much easier to follow the diagrams in the following link

    https://docs.microsoft.com/en-us/azure/machine-learning/concept-azure-machine-learning-architecture

    if they consist of ordered/numbered sequence of operations, e.g. training workflow, inference workflow.

    Thanks.

    1 vote
    Vote

    We're glad you're here

    Please sign in to leave feedback

    Signed in as (Sign out)
    You have left! (?) (thinking…)
    0 comments  ·  Flag idea as inappropriate…  ·  Admin →
  14. 1 vote
    Vote

    We're glad you're here

    Please sign in to leave feedback

    Signed in as (Sign out)
    You have left! (?) (thinking…)
    0 comments  ·  Flag idea as inappropriate…  ·  Admin →
  15. Enable schema configuration for partition column

    When creating a new tabular dataset (e.g. from parquet files which are partitioned into folders by date) then there is the option in the AML Studio to derive a column from the folder structure (e.g. '/date={date}/*.parquet).

    However this column is not included in the schema configuration in the next steps, so there is no chance to change it from 'String' to 'Date' for instance.

    1 vote
    Vote

    We're glad you're here

    Please sign in to leave feedback

    Signed in as (Sign out)
    You have left! (?) (thinking…)
    0 comments  ·  Flag idea as inappropriate…  ·  Admin →
  16. Support more shortcuts in Notebooks

    There are a number of Jupyter Shortcuts besides shift+enter that are frequently used, like X, A and B...

    Please add support to these shortcuts or hotkeys. Same for Synapse Analytics Notebooks.

    1 vote
    Vote

    We're glad you're here

    Please sign in to leave feedback

    Signed in as (Sign out)
    You have left! (?) (thinking…)
    0 comments  ·  Flag idea as inappropriate…  ·  Admin →
  17. Create AML Datasets from within the Designer

    If the Designer could create fully functional AML Dataset (Tabular and/or File), it would be possble to have citizen data scientists work on data transformation, preparation and feature engineering and then use AutoML or the SDK to do actual Machine Learning with these Datasets.

    At the moment there is a disconnect between Designer and other authoring methods.

    1 vote
    Vote

    We're glad you're here

    Please sign in to leave feedback

    Signed in as (Sign out)
    You have left! (?) (thinking…)
    1 comment  ·  Flag idea as inappropriate…  ·  Admin →
  18. Labelling project, does not incrementally add new data points

    I have updated the source dataset. Confirmed dataset version number is updated in labeling project. But added data in new version is not being updated in labeling project - 24h has past. (Should be an option to manually update the data within the labeling project instead of waiting 24h too)

    1 vote
    Vote

    We're glad you're here

    Please sign in to leave feedback

    Signed in as (Sign out)
    You have left! (?) (thinking…)
    0 comments  ·  Flag idea as inappropriate…  ·  Admin →
  19. Japanize Explainer on Notebook

    I would like to see Explainer available in Japanese on Notebook as well as the browser version.

    Our Japanese customer is considering implementing an automated ML service, and Azure ML is one of the candidates.
    However, the problem is that the Explainer cannot be displayed in Japanese on the Notebook, as shown in the attached image.

    1 vote
    Vote

    We're glad you're here

    Please sign in to leave feedback

    Signed in as (Sign out)
    You have left! (?) (thinking…)
    0 comments  ·  Flag idea as inappropriate…  ·  Admin →
  20. The Azure Machine Learning VS Code Extension does not work for me

    I use the App Service, Functions, and Storage extensions without any trouble. I have never gotten the Machine Learning extension to work properly.

    I have tried to debug multiple times with no success

    1 vote
    Vote

    We're glad you're here

    Please sign in to leave feedback

    Signed in as (Sign out)
    You have left! (?) (thinking…)
    0 comments  ·  Flag idea as inappropriate…  ·  Admin →
  • Don't see your idea?

Machine Learning

Categories

Feedback and Knowledge Base