My work has two long-term threads: using Efficient ML to identify the necessary structures that intelligence truly depends on, and using AI4Human to understand the human brain and expand the interaction bandwidth between carbon-based and silicon-based intelligence. I am also interested in other generative models that feel intellectually alive.
I treat compression as a scientific probe. Quantization, sparsity, and low-rank structure are not only engineering tools; they reveal which parts of a model remain essential after redundancy is stripped away. Those surviving structures are clues for next-generation AI architectures.
I want to use AI to understand the human brain, explore next-generation human-computer interaction, and increase the communication bandwidth between carbon-based and silicon-based intelligence.
I am interested in other compelling generative models as well, especially when they reveal new ways to sample, reason, or organize uncertainty.
Kaicheng Yang*, Xun Zhang*, Haotong Qin, Yucheng Lin, Kaisen Yang, Yulun Zhang
Xun Zhang*, Kaicheng Yang*, Hongliang Lu, Haotong Qin, Yong Guo, Yulun Zhang
Kai Liu*, Kaicheng Yang*, Zheng Chen, Zhiteng Li, Yong Guo, Wenbo Li, Linghe Kong, Yulun Zhang
Kaicheng Yang, Kaisen Yang, Baiting Wu, Xun Zhang, Qianrui Yang, Haotong Qin, He Zhang, Yulun Zhang
I am working on it.
Kaisen Yang, Jayden Teoh, Kaicheng Yang, Yitong Zhang, Alex Lamb