About me
I’m a 1st year PhD student at University of Edinburgh working on speech and language processing, supervised by Prof. Peter Bell. Prior to that I had obtained an MPhil degree at the Chinese University of Hong Kong (CUHK), supervised by Prof. Xunying Liu and Prof. Helen Meng, during which I focused on Automatic Neurocognitive Disorder Detection. I gained MEng and BA degrees in Information Engineering from the University of Cambridge.
Publications
Exploiting prompt learning with pre-trained language models for Alzheimer's Disease detection. | [Paper] Wang Y., Deng J., Wang T., Zheng B., Hu S., Liu X., Meng H., Accepted by ICASSP 2023 Exploring linguistic feature and model combination for speech recognition based automatic AD detection | [Paper] Wang Y., Wang T., Ye Z., Meng L., Hu S., Wu X., Liu X., Meng H., Proc. Interspeech 2022 Conformer Based Elderly Speech Recognition System for Alzheimer's Disease Detection | [Paper] Wang T., Deng J., Geng M., Ye Z., Hu S., Wang Y., Cui M., Jin Z., Liu X., Meng H., Proc. Interspeech 2022 Exploiting Visual Features Using Bayesian Gated Neural Networks for Disordered Speech Recognition. | [Paper] Liu S., Hu S., Wang Y., Yu J., Su R., Liu X., Meng H. ISCA Student Paper Award Nomination. Proc. Interspeech 2019 Exploring Self-Supervised Pre-Trained ASR Models For Dysarthric and Elderly Speech Recognition. | [Paper to be updated] Hu S., Xie X., Jin Z., Geng M., Wang Y., Cui M., Deng J., Liu X., Meng H., Accepted by ICASSP 2023
Education
University of Edinburgh, 2023 - ? Working for PhD degree at the Institute for Language, Cognition and Computation Chinese University of Hong Kong (CUHK), 2021 - 2023 MPhil degree at the SEEM Department University of Cambridge, Christ’s College, 2017 - 2021 BA and MEng Hons in Information and Computer Engineering
Research Experience
Neurocognitive Disorder (NCD) Detection, 2021-2023 • Used pre-trained models as text encoders and fed text features into an ensemble of back-end classifiers to perform automatic AD detection. • Used prompt-based fine-tuning paradigm to avoid pipeline system of text encoders and separated classifier, as well as make the fine-tuning objective consistent with the pre-training objective, which reduce the model performance variance • Now working on transferring the methods across language (applying the methods on Cantonese NCD detection data) and involving tighter integration between the ASR and AD detection components Multimodal Emotion Recognitions, Oct 2020 - May 2021 • Used the visual, acoustic and language (text) information to do emotion recognition for speaking videos Audio-visual Speech Recognition, Jul - Sep 2018 • Conducted image pre-processing for a disordered speech recognition task with audio-visual features.
Extra-curricular Activity
Cambridge University Women’s Basketball Club, second team Player, 2017 - 2020