Hi, I am Mingyu Cao (曹明宇), a first-year PhD student in the School of Computer Science and Electronic Engineering at the University of Surrey, advised by Dr. Lu Yin. My research focuses on Diffusion Language Models and Efficient LLMs.
Prior to my PhD, I worked as an NLP Algorithm Engineer at several leading Chinese internet companies, including NetEase, ByteDance, and Shopee, where I specialised in machine translation. Earlier, during my master's studies, my research focused on biomedical information extraction and knowledge-graph-based question answering, advised by Dr. Ling Luo.
Outside of research, I enjoy movies, social media, and travelling. I am a strong advocate for work-life balance. I am always happy to chat about research, life, travelling, or anything interesting.
Mingyu Cao, Alvaro H.C. Correia, Christos Louizos, Shiwei Liu#, Lu Yin# (# corresponding author)
ICML 2026 Regular
SOAR is a training-free decoding algorithm for Diffusion Language Models that adaptively switches between wider search and parallel decoding based on model confidence, improving reasoning and code generation quality without sacrificing inference speed.
Mingyu Cao, Alvaro H.C. Correia, Christos Louizos, Shiwei Liu#, Lu Yin# (# corresponding author)
ICML 2026 Regular
SOAR is a training-free decoding algorithm for Diffusion Language Models that adaptively switches between wider search and parallel decoding based on model confidence, improving reasoning and code generation quality without sacrificing inference speed.
Mingyu Cao, Gen Li, Jie Ji, Jiaqi Zhang, Ajay Jaiswal, Li Shen, Xiaolong Ma, Shiwei Liu, Lu Yin# (# corresponding author)
TMLR 2026
A pruning method for Mixture-of-Experts models that merges multiple experts per layer into a reduced set, preserving model quality while reducing memory usage and improving inference efficiency.
Mingyu Cao, Gen Li, Jie Ji, Jiaqi Zhang, Ajay Jaiswal, Li Shen, Xiaolong Ma, Shiwei Liu, Lu Yin# (# corresponding author)
TMLR 2026
A pruning method for Mixture-of-Experts models that merges multiple experts per layer into a reduced set, preserving model quality while reducing memory usage and improving inference efficiency.
Ling Luo, Zhihao Yang, Mingyu Cao, Lei Wang, Yin Zhang, Hongfei Lin
Journal of biomedical informatics (JBI) 2020-03-01
Joint extraction of entities and relations for biomedical text.
Ling Luo, Zhihao Yang, Mingyu Cao, Lei Wang, Yin Zhang, Hongfei Lin
Journal of biomedical informatics (JBI) 2020-03-01
Joint extraction of entities and relations for biomedical text.