Seminar 3 : Paul Michel
Title : Introductory class on transformers
Abstract : I will give a quick introduction on the transformer, a neural network architecture that has become ubiquitous in natural language processing since its introduction in 2017. I will describe the artchitecture itself, why and how it came to be and describe some of the research looking into why it works so well.
Centre Sciences des Données
de l’Ecole Normale Supérieure