Decision-Theoretic Safety Assessment of Persona-Driven Multi-Agent Systems in O-RAN
📰 ArXiv cs.AI
arXiv:2604.09682v1 Announce Type: cross Abstract: Autonomous network management in Open Radio Access Networks requires intelligent decision making across conflicting objectives, yet existing LLM based multi agent systems employ homogeneous strategies and lack systematic predeployment validation. We introduce a persona driven multi agent framework where configurable behavioral personas structured specifications encoding optimization priorities, risk tolerance, and decision making style influence
DeepCamp AI