At the moment the procedure for web service publication is not easy and intuitive. Most of time some Project Columns modules are needed after and before Score module in order to tell what column is target (response) and what columns in dataset must be considered as input. It must be easier in a visual surface than the programming environment like R to do this. But at the moment it is not the case with AzureML. After the recent update to publishing web service procedure, the result web service considers all the columns of input dataset as input parameter! So my suggestion is to 1) solve this issue and 2) make it easier for publishing a web service out of a model.
The reason that many people need something like AzureML is not a visual surface but ease of deployment.
You can use the ILearnerDotNet input and output parameters to do this. Thank you for your excellent product.
At the moment the procedure for web service publication is not easy and intuitive. Most of time some Project Columns modules are needed after and before Score module in order to tell what column is target (response) and what columns in dataset must be considered as input. It must be easier in a visual surface than the programming environment like R to do this. But at the moment it is not the case with AzureML. After the recent update to publishing web service procedure, the result web service considers all the columns of input dataset as input parameter! So my…3 votescompleted · Adminmluservoice (Program Manager, Microsoft Azure Machine Learning, Microsoft Azure) responded
We have overhauled the web service publishing workflow. Let us know what you think!
I'm enjoying Azure ML and the samples provided. After the 30 day trial do the samples remain or to we loose them?
Also, after the 30 day free trial, what changes? Do we still have the account and all the existing experiments? Then do we just have to pay for running experiments?1 votecompleted · Adminmluservoice (Program Manager, Microsoft Azure Machine Learning, Microsoft Azure) responded
The free tier has no end date. You can convert a trial into a paid workspace.
Have the option to choose which workspace to save an experiment to as well as giving it a name. Currently the only way to move an experiment is to cut and paste all the objects in it into an empty experiment in another workspace.
thx1 votecompleted · Adminmluservoice (Program Manager, Microsoft Azure Machine Learning, Microsoft Azure) responded
This is now available in production. Let us know what you think!
Thank you for the samples. How do we get more info about the sample data? There seems to be a Sample Section missing from ML Home page or the Info links are missing directly from the sample in Studion.1 vote
As far as I can tell there is no way to retrieve logs for runs triggered from web services.
Having them alongside the ML Studio executions in the run history would be great (a column indicating whether the run was triggered from ML studio or from a web service would be nice as well).
Another possibility (or even better - in addition) would be to dump the logs into a blob storage container in the associated Azure account, similar to Azure Website logs.5 votes
We have enabled logging which can be turned on in the web service configuration page.
At the end of the tutorial for the German Credit Risk data, I couldn't locate the "Ready For Production?" button. Did I miss something, was it moved?
I love this site. Thank you for the opportunity to practice.1 vote
This is no longer required – Here is revised tutorial – http://azure.microsoft.com/en-us/documentation/articles/machine-learning-publish-a-machine-learning-web-service/
It looks like Matchbox recommender doesn't support online learning at the moment. This makes impossible to provide recommendation for the new users based on their profile/preferences.44 votes
We have support for Vowpal Wabbit which provides online learning. We look for more opportunities based on customer scenarios and ask. We also support retraining capability which can be programatically triggered for ongoing model updates
Most financial data is available on the Internet - this includes equities, derivatives prices, economic data etc. All of these are in different formats and don't always come in neat CSV formats.
What we need in Azure ML is the freedom to read plain text, HTML or other formats as is. We can write the R Code to parse it into datasets internally. This will hugely improve the usability of the platform.5 votes
It is now possible to connect to internet from R module.
It would be great if we are allowed to upload a ZIP file containing a TXT or CSV dataset. This will save on bandwidth and time for end users. Currently I can upload a ZIP file but I am unable to visualize that as a normal dataset.58 votes
You can upload a ZIP file and then use an Unpack ZIP module to extract the dataset.
Use machine learning to redirect an incoming voice call (with speech recognition) to a self-service portal instead of an agent1 vote
This particular solution is not in our roadmap to be developed by our team at the moment. However, the users of the community are always adding new experiments so please check back often!
Looking though the new portal, there are many cool things. However, I haven't been able to find a way to access my AzureML workspaces from the new portal29 votes
Azure ML workspace can now be created in Ibiza
Allow getting the Original Experiment from the trained model.
After you saved train model as a trained model, you can delete/change the original experiment.
Being able to see the original Experiment the model is based on (with the right permissions), would help retain the knowledge.
On the same issue, if the original experiment exist, then some indication on the train model was saved as Trained Model similar to set/unset publish input/output.5 votes
This should be completed and available in the run history.
Currently only built-in models in Azure ML can be saved and published. We've built a model in R and want to publish it as well but Azure ML does not support it right now. I wonder if Azure ML can somehow put some rules on the input/output of the R-Script, so when a script follows these rules, it can be saved and published like any other built-in models. This could increase the extensibility and flexibility of Azure ML pretty much.120 votes
You can use create R model described here – https://msdn.microsoft.com/en-us/library/azure/dn955435.aspx
When building a model, running it takes quite some time and while working on it, I may not have fully specified all elements of my model as I create the outline of the proces.
Therefore I would like to be able to run the model up to a certain step, ignoring all elements after the selected block. This would save compute and wait-time and would fit my way of working nicely.61 votes
Please see the “Run Selected” feature in the Azure ML Studio which allows you to run only a group of selected modules within a graph.
I have just signed up for a trial of Azure Machine Learning, yet your site keesp telling me either: 1. My account does not exist, or 2. That I do not have access to the ML workspace I just created (with the account that apparently does not exist!).
A typically clunky Microsoft experience, I am afraid...1 vote
Nathaniel: hopefully you have been able to resolve this issue. If not, please file an issue in the Azure ML Users Forum (https://social.msdn.microsoft.com/forums/azure/en-US/home?forum=MachineLearning) and an engineer will assist you.
It would be great to see a public competition such as http://www.kaggle.com/c/titanic-gettingStarted used as the training content for learning AzureML.6 votes
Significant improvements have been done on learning materials. Please continue to provide feedback
Consumption pricing is very clear on the website but it would be great to 'predict' budget needs. Please publish examples of costs expected or being experienced by preview customers as soon as you can.6 votes
It would be nice to see Annotation capability in Experiment. That will allow users to put notes with free form text explaining purpose of the module and how it works, etc..12 votes
Comments are now available in the experiment modules.
When experimenting, the ability to disable steps makes it easier to try different approaches. For example, I wanted to have different sets of columns as inputs to a regression. I have 4-5 different Project Column steps with different fields and swap the output to the inputs to select. However, I'd love to just disable them on a right-click menu item.111 votes
The Run Selected feature satisfy your requirement. Please select the modules you want to run and just click the “Run Selected” button.
Azure CXP Community
I'm not familiar with R and don't wish to invest in learning it, when Python packages such as Scipy, Numpy, Pandas and Scikit-Learn meet most of my needs. I wish I could combine the power of these with Azure ML's slick interface, and easy-to-deploy APIs41 votes
This is now supported – here is the sample experiment – http://gallery.azureml.net/Details/978fa971bca845f39ee07fbbec23143d
- Don't see your idea?