Changsheng Wang

Papers on Lattice

Total citations

Topics

Publication activitypapers/week, last 8 weeks

Research focus

Architecture Design (Transformers, SSMs, MoE) (1)Training Efficiency & Optimization (1)

Frequent co-authors

Yicheng Lang (1)Yihua Zhang (1)Zheng Zhang (1)Wotao Yin (1)

Papers (1)

Feb 19, 2026

Yicheng Lang +53w ago

Powering Up Zeroth-Order Training via Subspace Gradient Orthogonalization

Zeroth-order optimization gets a huge boost: ZO-Muon slashes query requirements by 75% while simultaneously improving accuracy in LLM and ViT fine-tuning.

Yicheng Lang, Changsheng Wang, Yihua Zhang +3

Architecture Design (Transformers, SSMs, MoE)Training Efficiency & Optimization

Search

Changsheng Wang

Publication activitypapers/week, last 8 weeks

Research focus

Frequent co-authors

Papers (1)