Merge pull request #79263 from adyada/release-notes_@@10June2019

AzureML SDK Release notes @@10 june2019

Author: ktoliver

articles/machine-learning/service/azure-machine-learning-release-notes.md

@@ -20,7 +20,41 @@ In this article, learn about the Azure Machine Learning service releases.  For a
 
 See [the list of known issues](resource-known-issues.md) to learn about known bugs and workarounds.
 
-## 2019-06-10 
+## 2019-06-10
+
+### Azure Machine Learning SDK for Python v1.0.43
+
++ **New features**
+  + Azure Machine Learning now provides first-class support for popular machine learning and data analysis framework Scikit-learn. Using [`SKLearn` estimator](https://docs.microsoft.com/python/api/azureml-train-core/azureml.train.sklearn.sklearn?view=azure-ml-py), users can easily train and deploy Scikit-learn models.
+    + Learn how to [run hyperparameter tuning with Scikit-learn using HyperDrive](https://github.com/Azure/MachineLearningNotebooks/blob/master/how-to-use-azureml/training/train-hyperparameter-tune-deploy-with-sklearn/train-hyperparameter-tune-deploy-with-sklearn.ipynb).
+  + Added support for creating ModuleStep in pipelines along with Module and ModuleVersion classes to manage reusable compute units.
+  + ACI webservices now support persistent scoring_uri through updates. The scoring_uri will change from IP to FQDN. The Dns Name Label for FQDN can be configured by setting the dns_name_label on deploy_configuration. 
+  + Automated machine learning new features:
+    + STL featurizer for forecasting
+    + KMeans clustering is enabled for feature sweeping
+  + AmlCompute Quota approvals just became faster! We have now automated the process to approve your quota requests within a threshold. For more information on how quotas work, learn [how to manage quotas](https://docs.microsoft.com/azure/machine-learning/service/how-to-manage-quotas).
+ 
+
++ **Preview features**
+    + Integration with [MLflow](https://mlflow.org) 1.0.0 tracking through azureml-mlflow package ([example notebooks](https://aka.ms/azureml-mlflow-examples)).
+    + Submit Jupyter notebook as a run. [API Reference Documentation](https://docs.microsoft.com/python/api/azureml-contrib-notebook/azureml.contrib.notebook?view=azure-ml-py)
+    + Public Preview of [Data Drift Detector](https://docs.microsoft.com/python/api/azureml-contrib-datadrift/azureml.contrib.datadrift?view=azure-ml-py) through azureml-contrib-datadrift package ([example notebooks](https://aka.ms/azureml-datadrift-example)). Data Drift is one of the top reasons where model accuracy degrades over time. It happens when data served to model in production is different from the data that the model was trained on. AML Data Drift detector helps customer to monitor data drift and sends alert whenever drift is detected. 
+
++ **Breaking changes**
+
++ **Bug fixes and improvements**
+  + RunConfiguration load and save supports specifying a full file path with full back-compat for previous behavior.
+  + Added caching in ServicePrincipalAuthentication, turned off by default.
+  + Enable logging of multiple plots under the same metric name.
+  + Model class now properly importable from azureml.core (`from azureml.core import Model`).
+  + In pipeline steps, `hash_path` parameter is now deprecated. New behavior is to hash complete source_directory, except files listed in .amlignore or .gitignore.
+  + In pipeline packages, various `get_all` and `get_all_*` methods have been deprecated in favor of `list` and `list_*`, respectively.
+  + azureml.core.get_run no longer requires classes to be imported before returning the original run type.
+  + Fixed an issue where some calls to WebService Update did not trigger an update.
+  + Scoring timeout on AKS webservices should be between 5ms and 300000ms. Max allowed scoring_timeout_ms for scoring requests has been bumped from 1 min to 5 min.
+  + LocalWebservice objects now have `scoring_uri` and `swagger_uri` properties.
+  + Moved outputs directory creation and outputs directory upload out of the user process. Enabled run history SDK to run in every user process. This should resolve some synchronization issues experienced by distributed training runs.
+  + The name of the azureml log written from the user process name will now include process name (for distributed training only) and PID.
 
 ### Azure Machine Learning Data Prep SDK v1.1.5