3 papers
cs.DC2026
KVServe: Service-Aware KV Cache Compression for Communication-Efficient Disaggregated LLM Serving
Zedong Liu, Xinyang Ma, Dejun Luo +9
LLMs are widely adopted in production, pushing inference systems to their limits. Disaggregated LLM serving (e.g., PD separation and KV state disaggregation) improves scalability a…
math.PR2026
Structure preservation and emergent dissipation in stochastic wave equations with transport noise
Chang Liu, Dejun Luo
We study nonlinear wave equations perturbed by transport noise acting either on the displacement or on the velocity. Such noise models random advection and, under suitable scaling…
math.PR2025
Scaling Limit and Large Deviation for 3D Globally Modified Stochastic Navier-Stokes Equations with Transport Noise
Chang Liu, Dejun Luo
We consider the globally modified stochastic (hyperviscous) Navier-Stokes equations with transport noise on 3D torus. We first establish the existence and pathwise uniqueness of th…