Add release notes for ML.NET 0.4 (dotnet#656)

docs/release-notes/0.4/release-0.4.md

@@ -0,0 +1,88 @@
+# ML.NET 0.4 Release Notes
+
+Today we are releasing ML.NET 0.4. During this release we have started
+exploring new APIs for ML.NET that enable functionality that is missing from
+the current APIs. We welcome feedback and contributions to the
+conversation (relevant issues can be found [here](https://github.com/dotnet/machinelearning/projects/4)). While the
+focus has been on designing the new APIs, we have also moved several
+components from the internal codebase to ML.NET.
+
+### Installation
+
+ML.NET supports Windows, MacOS, and Linux. See [supported OS versions of .NET
+Core
+2.0](https://github.com/dotnet/core/blob/master/release-notes/2.0/2.0-supported-os.md)
+for more details.
+
+You can install ML.NET NuGet from the CLI using:
+```
+dotnet add package Microsoft.ML
+```
+
+From package manager:
+```
+Install-Package Microsoft.ML
+```
+
+### Release Notes
+
+Below are some of the highlights from this release.
+
+* Added SymSGD learner for binary classification
+  ([#624](https://github.com/dotnet/machinelearning/pull/624))
+
+    * [SymSGD](https://arxiv.org/abs/1705.08030) is a technique for
+      parallelizing
+      [SGD](https://en.wikipedia.org/wiki/Stochastic_gradient_descent)
+      (Stochastic Gradient Descent). This enables it to sometimes perform
+      faster than existing SGD implementations (e.g. [Hogwild
+      SGD](https://docs.microsoft.com/en-us/dotnet/api/microsoft.ml.trainers.stochasticgradientdescentbinaryclassifier?view=ml-dotnet)).
+    * SymSGD is available for binary classification, but can be used in
+      multiclass classification with
+      [One-Versus-All](https://docs.microsoft.com/en-us/dotnet/api/microsoft.ml.models.oneversusall?view=ml-dotnet)
+    * SymSGD requires adding the Microsoft.ML.HalLearners NuGet package to your project
+    * The current implementation in ML.NET does not yet have multi-threading
+      enabled due to build system limitations (tracked by
+      [#655](https://github.com/dotnet/machinelearning/issues/655)), but
+      SymSGD can still be helpful in scenarios where you want to try many
+      different learners and limit each of them to a single thread. 
+    * Documentation can be found
+      [here](https://docs.microsoft.com/en-us/dotnet/api/microsoft.ml.trainers.symsgdbinaryclassifier?view=ml-dotnet)
+
+* Added Word Embeddings Transform for text scenarios
+  ([#545](https://github.com/dotnet/machinelearning/pull/545))
+
+    * [Word embeddings](https://en.wikipedia.org/wiki/Word_embedding) is a
+      technique for mapping words or phrases to numeric vectors of relatively low
+      dimension (in comparison with the high dimensional n-gram extraction).
+      These numeric vectors are intended to capture some of the meaning of the
+      words so they can be used for training a better model. As an example,
+      SSWE (Sentiment-Specific Word Embedding) can be useful for sentiment
+      related tasks.
+    * This transform enables using pretrained models to get the embeddings
+      (i.e. the embeddings are already trained and available for use).
+    * Several options for pretrained embeddings are available:
+      [GloVe](https://nlp.stanford.edu/projects/glove/),
+      [fastText](https://en.wikipedia.org/wiki/FastText), and
+      [SSWE](http://anthology.aclweb.org/P/P14/P14-1146.pdf). The pretrained model is downloaded automatically on first use.
+    * Documentation can be found
+      [here](https://docs.microsoft.com/en-us/dotnet/api/microsoft.ml.transforms.wordembeddings?view=ml-dotnet).
+
+* Improved support for F# by allowing use of property-based row classes ([#616](https://github.com/dotnet/machinelearning/pull/616))
+
+    * ML.NET now supports F# record types.
+    * The ML.NET samples repository is being updated to include F# samples as part of [#36](https://github.com/dotnet/machinelearning-samples/pull/36).
+
+Additional issues closed in this milestone can be found
+[here](https://github.com/dotnet/machinelearning/milestone/3?closed=1).
+
+### Acknowledgements
+
+Shoutout to [dsyme](https://github.com/dsyme),
+[SolyarA](https://github.com/SolyarA),
+[dan-drews](https://github.com/dan-drews),
+[bojanmisic](https://github.com/bojanmisic),
+[jwood803](https://github.com/jwood803),
+[sharwell](https://github.com/sharwell),
+[JoshuaLight](https://github.com/JoshuaLight), and the ML.NET team for their
+contributions as part of this release!