Sourav Dutta
Huawei Multilingual NLP Tech in HMS
Multi-lingual language models have become the de-facto backbone for diverse NLP applications like semantic search, text understanding, classification and question answering. However, the computational expense of model training, fine-tuning for individual downstream tasks, annotated training data requirements, and domain-specific semantic understanding has impacted the wide adoption of such models in industrial settings. In this talk, we explore how we at Huawei Research are tackling these challenges with scientific breakthroughs in novel language understanding frameworks and domain-aware chatbot platforms to provide users with improved quality-of-experience and enhanced informational assistance.
back to overview
Watch Recording
