Cantonese Speech

Cantonese speech research focuses on developing robust and accurate speech processing technologies for this widely spoken but under-resourced language. Current efforts concentrate on improving automatic speech recognition (ASR) and machine translation (MT) systems, often employing deep neural networks (DNNs), including transformer architectures and hybrid DNN-HMM models, and leveraging techniques like back-translation and data augmentation to address data scarcity. These advancements are crucial for bridging the technological gap for Cantonese speakers and enabling applications such as in-car assistants and educational tools for dyslexic students, while also contributing valuable insights into low-resource language processing for the broader NLP community.