Reproducible Research in Network Modeling

This document bundle contains recent research on networked AI and wireless systems (6G/LWAM), LLM serving and caching, distributed MoE inference at the edge, NPU scheduling, UAV formation MAC control, and learning-based spectrum allocation. Collectively these works enable higher performance and autonomy but introduce new security and safety attack surfaces: KV-cache reservation uncertainty (preemption/recomputation) and memory-exhaustion/availability attacks against LLM serving; model-serving and distributed-MoE substitution strategies that can enable model-quality attacks, evasion or data-exfi