[jira] [Commented] (SOLR-10717) Learning to rank: a query will fail if the feature vector is requested without providing external feature information parameters

["Christine Poerschke (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/SOLR-10717?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16017738#comment-16017738 ] 

Christine Poerschke commented on SOLR-10717:
--------------------------------------------

bq. 2) Return a proper response error

This would seem the most logical to me, assuming that any actual re-ranking also fails if required External Feature Info is missing.

bq. 1) Return only the features for which it was possible to compute the value (ignore, or set to NaN features for which we had errors)

This could be a possibility too, though I'd be concerned about
* increased complexity for the consumer of the feature vector since we'd be turning a _number_ into a _possibly-missing-number_
* feature extraction vs. feature use consistency i.e. if the feature vector contains possibly-missing feature values then to be consistent would the models also need to support possibly-missing feature values presumably

bq. 3) Add a parameter (ignoreEfiErrors) that the user can set to true if s/he wants 1) otherwise 2).
This would be my least favourite choice, not so much in an 'ignoring errors is bad' philosophical sense but mainly because _ignoreEfiErrors_ would be another parameter to be decided upon by the user.

Here's a fourth option:
4) The ValueFeature class already takes a _required_ parameter. Perhaps an alternative to _ignoreEfiErrors_ could be to add a _requireEfi_ parameter (defaulted to true) to features that use External Feature Info.


> Learning to rank: a query will fail if the feature vector is requested without providing external feature information parameters
> --------------------------------------------------------------------------------------------------------------------------------
>
>                 Key: SOLR-10717
>                 URL: https://issues.apache.org/jira/browse/SOLR-10717
>             Project: Solr
>          Issue Type: Bug
>      Security Level: Public(Default Security Level. Issues are Public) 
>    Affects Versions: 6.5.1
>            Reporter: Diego Ceccarelli
>            Priority: Minor
>
> In ltr some features can depend on External Feature Informations that have to be provided at query time. If we query solr only to retrieve the feature vectors for the documents (without doing reranking), and without providing all the external feature informations used in the feature store the query will fail. 



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

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@lucene.apache.org
For additional commands, e-mail: dev-help@lucene.apache.org