[mlpack] mlpack 3.4.0 released

[Ryan Curtin]

Hello everyone!

I hope your Tuesday or Wednesday is going swimmingly.

Today I used an almost-complete automated script to tag and release
mlpack 3.4.0.  This contains a lot of the cool work that has been done
as a part of GSoC, but not all of it (more on that later!).  Here's a
condensed version of the parts of the changelog that are
external-facing, since it got pretty long:

 - Go bindings are added and stable.  See
   https://github.com/mlpack/mlpack-go.  (Technically these were added
   in 3.3.2, but not announced.)

 - R bindings are added and stable.  See
   https://www.mlpack.org/doc/stable/r_documentation.html.

 - one_hot_encoding binding added.  See
   https://www.mlpack.org/doc/stable/cli_documentation.html#preprocess_one_hot_encoding.

 - Bayesian linear regression binding added.  See
   https://www.mlpack.org/doc/stable/cli_documentation.html#bayesian_linear_regression.

 - Renamed InformationGain to HoeffdingInformationGain in
   `methods/hoeffding_trees/information_gain.hpp` (#2556).

 - CMake fixes: don't compile Python bindings by default (#2575), show
   error when Python modules not found (#2568).

 - Added macro for changing stream of printing and warnings/errors
   (#2556).

 - Added Spatial Dropout layer (#2564).

 - Replace boost program_options dependency with CLI11 (#2459).

 - Issue warnings when metrics produce NaNs in KFoldCV (#2595).

 - Added Soft Actor-Critic, Categorical DQN, and N-step DQN (#2487,
   #2454, #2461).

 - Add Silhoutte Score metric and Pairwise Distances (#2406).

 - Use case-sensitive categories in ARFF loader (#2486, #2516).

 - Fix defeatist search for spill tree traversals (#2566, #1269).

 - Fix incremental training of logistic regression models (#2560).

I'm not actually sure how good of a job I did compressing the changelog,
since quite a lot has changed.  This summer has seen massive amounts of
refactorings and code changes, in part from GSoC.  Awesome work!

I mentioned earlier I'd come back to this---the mlpack repository isn't
the only place where things are happening.  Recently, lots of nice
examples and additional model functionality have gone into the models
and examples repositories:

 - https://github.com/mlpack/examples/

 - https://github.com/mlpack/models/

And don't forget, lots of things are happening with ensmallen too
(although they don't get announced on this list):

 - https://github.com/mlpack/ensmallen/

In the next handful of days, mlpack 3.4.0 will percolate through
different package managers and become widely available.

If you have any issues, don't hesitate to report them on Github:
https://github.com/mlpack/mlpack/issues/ or on the chat channels.  See
also http://www.mlpack.org/community.html for more information.

Thank you so much to everyone who has contributed for this release (and
past releases)!  I'm really excited to see what we make as time goes on.

Have a great one!

Ryan

-- 
Ryan Curtin    | "I'm just going to shoot you once!"
ryan at ratml.org |   - Joseph Dunn