[ 
https://issues.apache.org/jira/browse/OAK-6317?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Tommaso Teofili updated OAK-6317:
---------------------------------
    Description: 
{{LMSEstimator}} update rule implementation wrongly updates the weights taking 
the LMS cost value into account, this means that if the error is huge (and the 
cost quadratically grows with it) the weight could get an update value almost 
proportional to the LMS value (smoothed by learning rate alpha).
The correct update rule should be used which only uses the residual.

The side effect of this bug is that:
* cost may vary a lot 
* weights could not converge to a good solution

  was:
{{LMSEstimator}} update rule implementation wrongly updates the weights taking 
the LMS cost value into account, this means that if the error is huge the (and 
the cost quadratically grows with it) the weight could get an update value 
almost proportional to the LMS value (smoothed by learning rate alpha).
The correct update rule should be used which only uses the residual.

The side effect of this bug is that:
* cost may vary a lot 
* weights could not converge to a good solution


> LMSEstimator update amount depending on cost amount
> ---------------------------------------------------
>
>                 Key: OAK-6317
>                 URL: https://issues.apache.org/jira/browse/OAK-6317
>             Project: Jackrabbit Oak
>          Issue Type: Bug
>          Components: solr
>    Affects Versions: 1.6.1, 1.7.1, 1.4.16, 1.2.26
>            Reporter: Tommaso Teofili
>            Assignee: Tommaso Teofili
>             Fix For: 1.8
>
>
> {{LMSEstimator}} update rule implementation wrongly updates the weights 
> taking the LMS cost value into account, this means that if the error is huge 
> (and the cost quadratically grows with it) the weight could get an update 
> value almost proportional to the LMS value (smoothed by learning rate alpha).
> The correct update rule should be used which only uses the residual.
> The side effect of this bug is that:
> * cost may vary a lot 
> * weights could not converge to a good solution



--
This message was sent by Atlassian JIRA
(v6.3.15#6346)

Reply via email to