I will be giving another presentation at the upcoming ATA Annual Conference in New Orleans, ATA59, jointly in the SciTech and Language Technology tracks. The presentation will give an introduction to neural machine translation. My talk is preliminarily scheduled for the very last time slot on Saturday before the final keynote. I hope to see you there, despite the late hour!
Abstract
“The end of the human translator,” “nearly indistinguishable from human translation” – these and similar headlines have been used to describe neural machine translation (NMT). Most language specialists have probably asked themselves: How much of that is hype? How far can this approach to machine translation really go? How does it work? The presentation will examine one of the available open source NMT toolkits as an illustrative example to explain the underlying concepts of NMT and sequence-to-sequence models. It will follow in the same spirit as last year’s general introduction to neural networks, which is summarized in the accompanying handouts.
Handouts
I have just uploaded the handout for the presentation onto the ATA server. The material is a slightly updated version of my blog post on neural networks, which summarizes my presentation at ATA58. You can download the handout here.
Copyright secured by Digiprove © 2021 Carola F Berger