Are Polyglots Better than the Rest?
Published in
Jan 21, 2022
Testing the performance of a polyglot Named Entity Recognition (NER) model on anonymised legal texts
In his recent article in Towards Data Science, “How to Cover Up in Multiple Languages”, Slimmer AI machine learning engineer Borach Jansema follows-up on his previous NER article, “Unlocking Inclusivity Possibilities with Polyglot-NER”.
In his article, Borach measures how well a polyglot (multilingual) NER model can switch to other domain types by taking a polyglot model trained on a general domain and applying it to a legal domain.
The English bias prevalent in NER effectively leaves out a considerable part of the world from this important technological advancement, polyglot models offer the possibility of a bridge between languages.
