Mulilingual NLP

SL5CE022

This lecture covers three topics related to develop NLP model and tools for many languages: • Neural Machine Translation: resources, evaluation and models • Multilingual Models: using neural networks to develop language independent models • Developing models for rare and under-resourced languages : challenges and techniques raised by developing NLP and ASR models beyond the twenty or so for which there are large amounts of annotated data (annotation and model transfert, fine-tuning, …)