The Turku.AI February meetup is hosted by TurkuNLP at the University of Turku and is rather naturally NLP themed. The place is Agora and the time 18:00 on Wednesday the 19.2. We have two great speakers:
Dr. Sofie Van Landeghem
Building customizable NLP pipelines with spaCy
spaCy is an industrial-strength Natural Language Processing library which allows you to quickly build performant pipelines to tackle specific NLP challenges. It comes with pretrained Machine Learning models for 10 languages and has basic support for over 50 languages. In this talk, Sofie will showcase the usage of spaCy as well as give a peak preview on the upcoming spaCy v.3 which will give the user full control over the model internals via configuration files.
Sofie van Landeghem is a renowned NLP expert who is also working on developing the widely used SpaCy library. She is visiting TurkuNLP at the time and we are happy to have the chance to welcome her to Turku.AI as well.
Assoc. Prof. Sampo Pyysalo
University of Turku
Deep transfer learning for Finnish NLP
Recently, transfer learning methods based on Google's transformer neural network architecture have notably advanced the state of the art in a range of natural language processing tasks. The Turku NLP group recently introduced FinBERT, a version of the transformer-based BERT model trained from scratch for Finnish. This talk introduces the process used to create FinBERT, presents advances using the model on a range of tasks including named entity recognition and syntactic analysis, and discusses future steps in deep learning for Finnish NLP.
Sampo Pyysalo has a long track record in language and speech technology. After working in Paris, Tokyo and Cambridge among others, Sampo has returned Turku as a fresh professor. He is one of the main characters behind FinBERT.