[jira] [Commented] (OPENNLP-1009) Experiment with deep learning based algorithms in OpenNLP

["Tommaso Teofili (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/OPENNLP-1009?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16068132#comment-16068132 ] 

Tommaso Teofili commented on OPENNLP-1009:
------------------------------------------

work is being done at https://github.com/apache/opennlp-sandbox/tree/master/opennlp-dl

> Experiment with deep learning based algorithms in OpenNLP
> ---------------------------------------------------------
>
>                 Key: OPENNLP-1009
>                 URL: https://issues.apache.org/jira/browse/OPENNLP-1009
>             Project: OpenNLP
>          Issue Type: Task
>            Reporter: Tommaso Teofili
>
> It'd be interesting to experiment with deep learning algorithms in OpenNLP.
> Possible experiments can include using recurrent and / or convolutional neural networks for NER, language detection, document categorization, etc.
> We could just leverage such algorithms during the training phase and let the prediction phase as it is, as ti minimize impact on API / users.
> From the tooling perspective we could start experimenting with DeepLearning4J and Nd4j.



--
This message was sent by Atlassian JIRA
(v6.4.14#64029)