Jingyang Xiang (ι‘Ήιι³)
CVByteDance Seed
ByteDance Seed
I work as an Algorithm Engineer at the Seed Speech Team of ByteDance. My research focus lies in building efficient large language models (LLMs).
I obtained my M.S. degree from the Advanced Perception on Robotics and Intelligent Learning Lab (APRIL), Institute of Cyber-Systems and Control, College of Control Science and Engineering, Zhejiang University (ZJU) in 2025, under the supervision of Prof. Zaisheng Pan and Prof. Yong Liu. Prior to that, I obtained my B.S. degree in Electrical Engineering and Automation from Zhejiang University of Technology in 2022, with Prof. Qi Xuan as my supervisor.
DFRot: Achieving Outlier-Free and Massive Activation-Free for Rotated LLMs with Refined Rotation.
Conference on Language Modeling (COLM 2025)
OvSW: Overcoming Silent Weights for Accurate Binary Neural Networks.
European Conference on Computer Vision (ECCV 2024)
MaxQ: Multi-Axis Query for N:M Sparsity Network.
IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR 2024)
SUBP: Soft Uniform Block Pruning for 1xN Sparse CNNs Multithreading Acceleration.
Conference on Neural Information Processing Systems (NeurIPS 2023)
Journal Reviewer: IEEE TCSVT, etc.
Conference Reviewer: NeurIPS 2024, etc.