Knowledge enhancement for pretrained LM & Hebrew NLP - it's all about the data
Details
We are happy to present out next Meetup session, with two great talks about Knowledge enhancement and Hebrew NLP .
Agenda (Wednesday, February 17th):
19:30-20:15 - Ohad Rozen - Entrepreneur, a PhD student at Bar Ilan University
Knowledge enhancement for pretrained language models
Large pretrained models such as GPT and BERT have significantly improved the state-of-the-art for a wide range of NLP tasks. Yet, they do not contain any explicit grounding to real world entities and as a result have difficulty recovering factual knowledge (e.g. ‘pangolin’ is an animal, ‘Baytown‘ is in Texas). We will review a few knowledge injection methods and present our approach (and architecture) for injecting facts into BERT in real time.
20:15 -20:45 - Amit Man, CTO DigitalOwl
Hebrew NLP - it's all about the data
In this talk we will look into DigitalOwl use case - medical data extraction for Israelis' Insurance companies, and look what happen when advance NLP models have access to Hebrew large corpus and tagged data
You are all invited to join the NLP IL community at: https://www.facebook.com/groups/2752846491451584 And https://www.linkedin.com/groups/13834904/