MLMI 6 Advanced Speech Recognition

Aims

The aim of this module is to teach students techniques used in advanced automatic speech recognition (ASR) systems with a focus on end-to-end (E2E) trainable models, large foundation models and the use of other large models that can be integrated into ASR systems. 

Objectives:

On completion of this module, students should:

• understand techniques used in state-of-the-art speech recognition systems 
• understand the relationship between E2E systems and classical ASR systems
• understand the use of large pre-trained models in ASR systems.