This talk gives an overview of the Flair Framework for Natural Language Processing (NLP). It's two main features are that (1) it is very easy to use and (2) that it gets state-of-the-art accuracies on many NLP tasks. I'll give an introduction from the practitioner's side and show how Flair can be used for tasks such as Named Entity Recognition or Sentiment Analysis on your data, and show how you can train your own models. I'll also present our recent research in Few-Shot and Zero-Shot Learning, and introduce an approach that can (1) learn new tasks with few training examples, (2) learn new tasks with no training examples at all and (3) operate in a "Continual Learning" setting in which it sequentially learns more and more tasks - without forgetting the old tasks.
About Alan Akbik
Alan joined the Humboldt-Universität zu Berlin in 2020 as professor of machine learning. His group focuses on natural language processing (NLP), i.e. methods that enable machines to understand human language. His research is made available in form of the open source NLP framework Flair that allows anyone to use state-of-the-art NLP methods in their research or applications. Before that, he spent many years in industrial research labs, first at IBM Research in California, then at Zalando Research in Berlin.