When training an entity recognizer on comprehend using a entity list of abbreviation. Would it start identifying other abbreviations not on the list?

Question

Basically: If I train a custom entity recogniser on comprehend using an entity list of gene abreviations, would it potentially start identifying other abbreviations that weren't in the entity lists as gene names?

The model has identified the abbreviation UVB as a gene name, however it's not in any of my dictionaries and I'm not sure why this would have happened

Answer 1

Correct, Comprehend's entity list mode is an easy way to provide a non-exhaustive list of entities while training a model. The model will learn the context in documents with these entities and subsequently generalize to unseen entities in similar context.