[jira] [Commented] (SPARK-14200) The optimization method of convex function

["Nick Pentreath (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/SPARK-14200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15215715#comment-15215715 ] 

Nick Pentreath commented on SPARK-14200:
----------------------------------------

I would advise first implementing this as a Spark Package (http://spark-packages.org/). See e.g. https://github.com/databricks/spark-tfocs. If it gains wide user adoption it could then be considered for inclusion in MLlib.

> The optimization method of convex function 
> -------------------------------------------
>
>                 Key: SPARK-14200
>                 URL: https://issues.apache.org/jira/browse/SPARK-14200
>             Project: Spark
>          Issue Type: Question
>          Components: MLlib, Optimizer
>    Affects Versions: 2.1.0
>            Reporter: chenalong
>              Labels: BMRM, MLlib, Optimization
>   Original Estimate: 1,344h
>  Remaining Estimate: 1,344h
>
> I want to implement Bundle Methods for Regularized Risk Minimization(BMRM) in Spark MLlib. BMRM is a nonsmooth convex optimization techniques, which is more faster than SGD and can solve non-differentiable problems and differentiable problems. Is this idea OK, Can you give me some advices?



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@spark.apache.org
For additional commands, e-mail: issues-help@spark.apache.org