Rethinking Sampling Strategy in Link Prediction

This arXiv feed contains multiple papers across network science, social simulation, and applied systems. Notable items with security implications: (1) "Simulation of Language Evolution under Regulated Social Media Platforms" (LLM + genetic algorithm) demonstrates realistic iterative content-evasion strategies that can bypass automated moderation—high misuse potential for policy evasion and illicit marketplaces; (2) "Toward Temporal Realism in City-Scale Crisis Response Simulation using LLM Agents" shows LLM-based social simulators lack realistic timing and recommends decoupling "when" (self-ex