[
https://issues.apache.org/jira/browse/MATH-1656?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17741665#comment-17741665
]
Gilles Sadowski commented on MATH-1656:
---------------------------------------
bq. [...] quite bulky dependency [...]
No, the API is separate from the implementation(s). Cf. link in the previous
comment.
bq. [...] a lot of I/O [...]
Not unless the user asks for it. For debugging, where one might want to see
messages at the lower levels ("DEBUG" and "TRACE"), poorer performance (due to
copious logging) is not an issue.
bq. CPU intensive applications
At the other end of the usage spectrum, configuring the level to "ERROR" will
trigger I/O only for critical problems, usually followed by an exception.
bq. Embedded applications
The framework allows for the implementation to be "no-op".
bq. [...] remove the DebugMode option and store the trajectory all the time.
It's a possibility (TBD) but then, it should be included/used by all optimizers.
In the context of this report, we should focus on what is strictly necessary
for the the new algorithms to provide the exact same service as the existing
ones. Each additional feature should be discussed separately (preferably on
the "dev" ML) and have a new JIRA associated to them.
> Classical multivariate optimizers (gradient descent, Raphson-Newton, BFGS)
> are missing
> --------------------------------------------------------------------------------------
>
> Key: MATH-1656
> URL: https://issues.apache.org/jira/browse/MATH-1656
> Project: Commons Math
> Issue Type: Wish
> Components: legacy
> Affects Versions: 4.0-beta1
> Reporter: François Laferrière
> Priority: Major
> Labels: features
> Attachments: MATH-1656-GradientDescent-Newton-BFGS-v2.0.zip,
> MATH-1658-GradientDescent-Newton-BFGS.patch, Screenshot from 2023-07-10
> 12-13-38.png
>
>
> Some classical multivariate such as
> * gradient descent,
> * Raphson-Newton,
> * BFGS
> are missing.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
