Handling Variable Element Presence in Parsing Document Task

[abdelaziz-jaddi]

Hello,

Sometimes, when tackling a document comprehension problem, it can be observed that elements are not always consistently present in documents, as illustrated by this example:

Document 1 (element 1: present, element 2: present, element 3: absent)

Document 2 (element 1: absent, element 2: present, element 3: present)

To prepare the training data, it is possible to add the category "other" when the element is not present in the document ?