I am Xinyu Yang (杨新宇), a first-year Ph.D. student at Carnegie Mellon University, advised by Prof. Beidi Chen. My research focuses on building useful, reliable, and efficient foundation model systems for scientific research and broader human society. I also work closely with Prof. Tianqi Chen at CMU and Prof. Huaxiu Yao at UNC.

Before coming to CMU, I was an undergraduate student from ACM Honors Class, Zhiyuan College, Shanghai Jiao Tong University, and a research intern working with Prof. Junchi Yan at SJTU-ReThinkLab. I also had a wonderful time as a research intern working with Prof. Song Han at MIT HAN Lab and remotely with Prof. Chelsea Finn at Stanford IRIS Lab.

Education

  Carnegie Mellon University
  Ph.D. student in ECE, advised by Prof. Beidi Chen. Aug. 2023 - Present

  Stanford University
  Research Intern, advised by Prof. Chelsea Finn. Mar. 2022 - Present

  Massachusetts Institute of Technology
  Research Intern, advised by Prof. Song Han. Nov. 2021 - Jun. 2023

  Shanghai Jiao Tong University （Ranking: 1/29, GPA: 4.0/4.3）
  B.Eng. in CS, advised by Prof. Yong Yu and Junchi Yan. Sep. 2019 - Jun. 2023

Recent Publications

$\mit{Preprint}$
[1] Harry Dong, Xinyu Yang, Zhenyu Zhang, Zhangyang Wang, Yuejie Chi, Beidi Chen, Get More with LESS: Synthesizing Recurrence with KV Cache Compression for Efficient LLM Inference, arXiv 2402.09398. [arXiv]

[2] Chenhang Cui*, Yiyang Zhou*, Xinyu Yang, Shirley Wu, Linjun Zhang, James Zou, Huaxiu Yao, Holistic Analysis of Hallucination in GPT-4V(ision): Bias and Interference Challenges, arXiv 2311.03287. [arXiv]

$\mit{2024}$
[1] Huaxiu Yao*, Xinyu Yang*, Xinyi Pan, Shengchao Liu, Pang Wei Koh, Chelsea Finn, Leveraging Domain Relations for Domain Generalization, in Proceeding of the 12th International Conference on Learning Representations (ICLR 2024), Vienna, Austria, May 2024 (Spotlight). [arXiv]

$\mit{2023}$
[1] Xinyu Yang*, Huaxiu Yao*, Allan Zhou, Chelsea Finn, Multi-Domain Long-Tailed Learning by Augmenting Disentangled Representations, Transactions on Machine Learning Research (TMLR), 2023. [arXiv]

[2] Zhijian Liu*, Xinyu Yang*, Haotian Tang, Shang Yang, Song Han, Flattened Window Attention for Efficient Point Cloud Transformer, in The IEEE/CVF Conference on Computer Vision and Pattern Recognition 2023 (CVPR 2023), Vancouver, Canada, Jun. 2023. [arXiv]

[3] Zhijian Liu*, Haotian Tang*, Alexander Amini, Xinyu Yang, Huizi Mao, Daniela Rus, Song Han, BEVFusion: Multi-Task Multi-Sensor Fusion with Unified Bird’s-Eye View Representation, in 2023 IEEE International Conference on Robotics and Automation (ICRA 2023), London, UK, Jun. 2023. [Project Page] [Code] [arXiv]

$\mit{2022}$
[1] Danning Lao*, Xinyu Yang*, Qitian Wu, Junchi Yan, Variational Inference for Training Graph Neural Networks in Low-Data Regime through Joint Structure-Label Estimation, in Proceedings of the 28th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (KDD 2022), Washington DC, Aug. 2022 (Research Track). [PDF] [Code] [Slides] [Poster]