Shizheng Zhao

Shanghai Jiao Tong University

Papers on Lattice

Total citations

Topics

h-index

Publication activitypapers/week, last 8 weeks

Research focus

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

Frequent co-authors

Shuyao Qi (1)Haoyuan Liu (1)Shizhen Zhao (1)

Papers (1)

Apr 21, 2026

Apr 21, 2026·also UConn

FEPLB: Exploiting Copy Engines for Nearly Free MoE Load Balancing in Distributed Training

Hopper's idle NVLink Copy Engine can be turned into a nearly free communication channel for MoE load balancing, slashing token stragglers by up to 70% without impacting existing parallelism.

Shuyao Qi, Haoyuan Liu, Shizheng Zhao +1

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

Search

Shizheng Zhao

Publication activitypapers/week, last 8 weeks

Research focus

Frequent co-authors

Papers (1)