[
https://issues.apache.org/jira/browse/SPARK-16365?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15396480#comment-15396480
]
Jay White Bear commented on SPARK-16365:
----------------------------------------
The discussion is about adding a full blown linear algebra library which most
ppl I've talked to think would be a great idea and aid in pipeline development
for ml projects, but it seems it's reached some sort of stasis here. It would
be redundant and it is not highly necessary in a programmatic sense- people can
use other libraries/packages. However, it would add value and streamline
delivery of complete packages for people who develop with Spark for other
applications, ease development for data scientists and other people who are not
primarily programmers, and as stated before make ml pipeline development a lot
more straightforward. I cant really speak to all your arguments, I just want
to build a EKF filter in Spark for SLAM and that requires basic matrix
operations. I need to do this in Spark right now. I don't need to create an
entire linear algebra yet, would you consider a pull request for matrix
addition and, then, if that goes well we can expand the discussion?
> Ideas for moving "mllib-local" forward
> --------------------------------------
>
> Key: SPARK-16365
> URL: https://issues.apache.org/jira/browse/SPARK-16365
> Project: Spark
> Issue Type: Brainstorming
> Components: ML
> Reporter: Nick Pentreath
>
> Since SPARK-13944 is all done, we should all think about what the "next
> steps" might be for {{mllib-local}}. E.g., it could be "improve Spark's
> linear algebra", or "investigate how we will implement local models/pipelines
> in Spark", etc.
> This ticket is for comments, ideas, brainstormings and PoCs. The separation
> of linalg into a standalone project turned out to be significantly more
> complex than originally expected. So I vote we devote sufficient discussion
> and time to planning out the next move :)
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org
For additional commands, e-mail: issues-h...@spark.apache.org
