Daisy Yu

Email: x.yu-99@sms.ed.ac.uk

Website: https://xiaorandaisyyu.github.io/

Research keywords: Speech Recognition, Speech Impairments

Bio:

Xiaoran (Daisy) is originally from China and has lived in the UK since 2021. She obtained her MEng degree in Electronic and Information Engineering from Imperial College London in 2025. During her time at Imperial, she developed an interest in speech and language processing. She was also a member of the Imperial College Sign Language Society, where she learned about the deaf/hard-of-hearing community. Her Master’s thesis focused on improving an existing automatic speech recognition system’s performance on hearing-impaired speakers’ speech. Her interest in inclusive speech technologies and aspiration in bridging the communication gap for the underrepresented has encouraged her to keep on researching in this area.

PhD research:

Speech produced by individuals with speech impairments (e.g., dysarthria) or hearing impairments (e.g., deafness) falls under the category of atypical speech. It differs from typical speech in various aspects, including intelligibility, acoustic properties, prosodic features, etc., posing challenges for current speech recognition systems. Furthermore, research in this area is often hindered by the limited availability of atypical speech data. Xiaoran aims to investigate methods that mitigate such performance degradation and data scarcity, with the broader goal of making speech technologies more accessible and explainable.

Supervisors: Hao Tang, Sharon Goldwater