Machine Learning
Welcome! You can use this site to tell the Azure Machine Learning team what features you would like to see.
Azure Machine Learning enables developers and data scientists to build advanced analytic web services in minutes and hours, eliminating the heavy lifting associated with deploying machine learning in modern data driven applications.
-
Time series models
Would love to see a time series prediction model using a MultiClass Neural Net. Also, where's the N# documentation?
173 votesplanned ·
Adminmluservoice
(Program Manager, Microsoft Azure Machine Learning, Microsoft Azure)
responded
We are planning to roll out custom modules in gallery which will have modules for timeseries such as https://gallery.cortanaintelligence.com/Experiment/Time-Series-Forecasting-using-Custom-Modules-1
-
Azure ML Integration with Tableau, Qlikview & PowerBI
Many visual IDEs for Machine Learning are fast growing today.
They provide a way a mechanism to directly process data and enable consumption of data in popular BI tools. This should be done for Azure ML as well73 votesplanned ·Adminmluservoice
(Program Manager, Microsoft Azure Machine Learning, Microsoft Azure)
responded
PowerBI is on the roadmap.
-
Configurable timeout for experiments and web services
Some bugs/errors could cause experiments or web service to run longer than expected. This can lead to high costs if several executions never ends.
There should be configurable timeouts, so that any run that exceeds this timeout is terminated:- Workspace global timeout: applies to all experiments and web services in a workspace
- Experiment and web-service specific timeouts62 votesplanned ·Adminmluservoice
(Program Manager, Microsoft Azure Machine Learning, Microsoft Azure)
responded
planned for web services.
-
Multiple inputs for Evaluate Model (only 2 currently possible)
Would be easier to work with & compare multiple models. having to cross-pairing all models is too slow to design & hard to evaluate.
Even if the output when multiple models would have fewer features than the default 2 input evaluate. (ex: fewer visualization options, metrics) But at least a good subset of main evaluation metrics should be visible.
57 votesplanned ·Adminmluservoice
(Program Manager, Microsoft Azure Machine Learning, Microsoft Azure)
responded
This feature is under planning.
-
Export Azure Machine Learning Studio's Workflow to Python/Anaconda codes.
I want to use the workflow at on premise Python/Anaconda environments for more speed, more customize, more cost-saving.
36 votes -
Pre-install custom R packages
I want to expose an R package I wrote as a web service. Since it's a custom R package they docs say the install has to happen every time the user user my service.
There's going to be a lot of repetitive overhead going this route, Any chance I can pre-install my package so the user only has to consume using the web service?
30 votesplanned ·Adminmluservoice
(Program Manager, Microsoft Azure Machine Learning, Microsoft Azure)
responded
We are planning to release a fix pretty soon (in week or 2) to address this issue
-
Visualization graphs - Exportable
I am trying to use the 'compare' feature in the visualization window, where it shows you graphs that compare two columns, but it is hard to see. The window isn't resizable and the axis-labels are too small/cut short so I can't actually see what it is I am analyzing. It would be perfect if you could export this graph -- even if to a JPEG or PDF, so you could visualize it better. Thank you!
23 votesplanned ·Adminmluservoice
(Program Manager, Microsoft Azure Machine Learning, Microsoft Azure)
responded
Thank you; the team will review as part of our planning process.
-
Add filter capability to reader
To avoid unnecessary resource consumption and get only those data needed from a table store please add a filter to the table store reader!
22 votes -
Add feature selection methods using classification/regression algorithms
Just for an example, I want to select features with classification/regression algorithms such as Multi Layer Perceptron, Linear regression, Decision tree and so on, selecting the best feature subset by performing forward/backward Best-first search.
This is can be done in WEKA.
20 votes -
18 votesplanned ·Adminmluservoice
(Program Manager, Microsoft Azure Machine Learning, Microsoft Azure)
responded
This feature is under review.
-
Save Python model objects as trained models
It would be nice to be able to save fitted objects from sklearn as trained models (similar to what is now doable with R scripts) without having to go through the process of pickling, reading a dataset, and unpickling.
13 votesplanned ·Adminmluservoice
(Program Manager, Microsoft Azure Machine Learning, Microsoft Azure)
responded
This feature is on our roadmap.
-
Add Power Query datasets published to the Corporate Data Catalog as a source in Azure Machine Learning
It would be fantastic if we could use Power Query datasets saved to the Corporate Data Catalog as a source in Azure Machine Learning.
Here is a blog post in which I describe it in more detail and outline the benefits. http://www.christoolivier.com/missed-opportunity-power-query-in-azure-machine-learning/
12 votesplanned ·Adminmluservoice
(Program Manager, Microsoft Azure Machine Learning, Microsoft Azure)
responded
This is currently being planned by the team. ETA is still pending.
-
Add possibility to extend timeout time for web service request
I'm trying to build a web service that does much more than just taking some already processed data and runs it through a model to determine some classification. This service should be able to take raw data, process to specified format, do some NLP computing on the data using trained databases and return the result to the user. When raw data is larger that around 200kb, the service does not have the time to finish before getting a timeout, and I have no ide why it cannot be extended more than 90s. It could be solved using batch request, but…
11 votes -
Boosted / bagged logistic regression
There is an ensemble learner for decision trees but not for logistic regression. Could this be added to ML studio?
9 votesplanned ·Adminmluservoice
(Program Manager, Microsoft Azure Machine Learning, Microsoft Azure)
responded
Adding additional ensemble learners is planned.
-
8 votesplanned ·Adminmluservoice
(Program Manager, Microsoft Azure Machine Learning, Microsoft Azure)
responded
Azure is planning to add a support for free offer that will make this easier
-
Performance statistics for Two-Class Logisitc Regression (missing R-squared, p-value)
Azure is great! However some features are still missing. For instance, for widely used Two-Class Logistic Regression there is no way (at least, I could not find any) to see the performance statistics, such as R-squared, p-value, likelihood ratio or any other. Those show the determinative power of model itself and bring very valuable information for assessment.
8 votesplanned ·Adminmluservoice
(Program Manager, Microsoft Azure Machine Learning, Microsoft Azure)
responded
We’re planning to add more model performance metrics.
-
Export to the R scripts from Azure ML workflows.
I want to use the workflow in Power BI's R script and SQL Server's R Server.
6 votesplanned ·Adminmluservoice
(Program Manager, Microsoft Azure Machine Learning, Microsoft Azure)
responded
we are looking for this scenario in future releases
-
It would be nice to have 'Logical Control Modules'. To be able to enroute or finish the experiment based on the previous modules output
This way user can branch out the module based on the output data from any previous module. This gives flexibility to end module if rest of the steps are not necessary.
7 votesplanned ·Adminmluservoice
(Program Manager, Microsoft Azure Machine Learning, Microsoft Azure)
responded
Support for control flows is on our roadmap.
-
Include Jaccard distance as a metric option
No distance measure is available for categorical features.
4 votesplanned ·Adminmluservoice
(Program Manager, Microsoft Azure Machine Learning, Microsoft Azure)
responded
This feature is being planned.
-
Simplify Deep Learning Experimental Setup
Net# is fairly intuitive, but I'm thinking of a more but all the stride/range/kernal stuff is way over the average human's head... in literature, we often see visualizations of the final arrangement used in some given Deep Learning model. Could you facilitate visually designing the model in the first place? (e.g. we just draw lines on the canvas and you fill in the numbers behind the scenes?)
1 vote
- Don't see your idea?