I use Watson Knowledge Studio to analyze resume. When I upload a document, a sentence boundary detection is run by Watson. However, a resume is not exactly like natural language such as emails or comments, and have less punctuation. Therefore, the sentence boundary detection can fail miserably, and split tokens that should be within the same entity on multiple lines.
To handle this problem, I created my own model to detect sentences in a resume. Now, I would like to upload the document to Watson, without letting it try to re-segment sentences. The best approach that I manage was two put two lines break each time my model predicted a sentence break. Thanks to that, Watson never joins together different sentences. However, it will sometimes consider that a sentence break is missing and add a new one.
How can I deactivate sentence boundary detection in Watson Knowledge Studio?
Unfortunately there is no option to disable/modify sentence boundary detection in Watson Knowledge Studio.