I do not see an Annotator Component page post migration in WKS. Is this a bug?


I am able to create regular expressions in WKS. But, I am unable to create a pre-annotator component that could apply this to documents.

The help documentation says – Log in as a Watson Knowledge Studio administrator and open the Annotator Component page.

However in the new WKS after moving to IBM Cloud, I do not see an Annotator Component page. Is this a bug?


Is it possible to integrate cloud API services provided from different country cloud center?

Is it possible to integrate cloud API services provided from different country cloud center?

Just like the title., I’m wondering that If I could integrate the services such as Natural Language Understanding and Watson Knowledge Studio.

What I want is that I Want to deploy the custom annotator generated from WKS into NLU so that I could get entities and relations as the language Korean is not a default supported language.

The problem is that I need to use NLU service from Korean cloud center and WKS service from US cloud center.

Is it possible?

As I know when one want to deploy the WKS M/L annotator into NLU, the one need to use the model ID(key or what ever) of the WKS M/L annotator so that the NLU could detect it.

And also the instances of NLU and WKS should be generated by same account(such as bluemix ID) because the link between NLU and WKS is performed inside the account instances.

So I wanna know the solution here about this problem.

Can I use different cloud center using one account?

Can I deploy the WKS M/L annotator generated by others account into NLU generated by mine?

Can I deploy the WKS M/L annotator into NLU by other way which is not using the model ID query and is using other method?

Can I deploy the WKS M/L annotator into NLU by using the WKS M/L annotator zip file that is exported from WKS?

Are there other ways?

Please let me know Thanks guys.


  • No Related Posts

Current state of BigInsights TextAnalytics – Building Datacap Annotators


Does anyone know what are the current state plans for replacing the BigInsight’s Text Analytics UI for building Datacap annotators? Every partner session and IBM site I’ve visiting list this product as being deprecated. If the product is deprecated, what are the plans to provide a UI for designing, developing, testing, and compiling annotators for Datacap? I’ve been building a significant amount of annotators for POCs and it’s difficult for me to communicate the future state of these components to customers. I hope I’m not the only one out there trying to find an answer to this question.


Watson Explorer Content Analytics Studio – How to modify annotator output?


My apologies if this is foolish question but so far I’ve been unable to find a way to modify the output of an annotator made in CA Studio. I know that this can be done when developing annotators manually with Apache UIMA but how can I do this with CA Studio?

Example of what I’m trying to accomplish:

Annotator: Margin_Search >
Input: “Margin for X was 5.5% per quarter” >
Desired output: 5.5%

Thank you


  • No Related Posts

WKS Machine Annotator: Problem indentifying a specific entity

I have created a machine annator / model based on about 50 very similarly structured documents. At this stage, when performing an evaluation on the scores, I have a persistant problem with identifying a entity TITLE – this is normaly a multi-word mention, enclosed in quotation marks and typically found three times in any one document. It is part fo thr text and not part of any header etc. All mentions of TITLE are correctly annotated in the current ground truth (I have checked them all!). All of the titles are also on my system as a dictionary. When running the Dictionary pre-annotator, ALL of the titles are correctly annotated…

However, when I run the machine annotator, the model only labels the first word or two as TITLE and then labels the entire rest of the metion as ORGANISATION. Literally ALL other entities are identified correctly.

I have reviewed all the help documentation but not managed to crack the problem. I have even tried to retrain without using my ORGANISATION dictionaries, but is still returns the same sort fo result whenever TITLE is identified.

Anyone got ANY suggestions?!


WKS+NLU vs Watson Explorer

I created a model in WKS and I can use it with NLU to extract entities. Can I do the same thing with Watson Explorer? Can I make a custom annotator in Content Analytics Studio (I know that we can export WKS model too) and apply this annotator to my collections?

Another question: Can we analyse one document (not a collection of documents) in Watson Explorer? Or WEX is for searching and analyzing the amount of documents?