Abstract:

Multi-agent large language models (LLMs) have become ubiquitous in applied AI, yet their theoretical foundations remain surprisingly understudied. Viewed through the lens of multi-agent systems theory, several shortcomings come to light: a lack of social intelligence, the absence of coordination mechanisms among agents, unknown emergent behavior, and interactions between agents that are bounded by natural language.

We address two of these gaps: the absence of social behavior and the lack of mechanisms for inter-agent coordination. We introduce Epistemic Probabilistic Language Agents (EPLA), a neuro-symbolic architecture for multi-agent coordination under uncertainty. A Symbolic Guard provides structured diagnostic feedback. The LLM generates typed actions, and the Guard controls their execution against an authoritative symbolic state. We formalize the epistemic layer in a gossip testbed through epistemic lottery gossip models, which combine view-based call histories with agent-indexed probability weights. We argue that implementing such a formalism can address shortcomings of agentic LLMs.

Bio:

Ebrahim Ardeshir-Larijani received his PhD in Computer Science from the University of Warwick and subsequently completed postdoctoral work at the University of Grenoble Alpes. He later joined the Institute for Research in Fundamental Sciences (IPM) as a senior postdoctoral fellow. He is currently an Assistant Professor of Computer Science at Iran University of Science and Technology (IUST). He also heads the IUST Centre for Information and Communication Technology (IT Service) and is a part-time researcher at the Pasargad Institute for Advanced Innovative Solutions. His research interests range from computational logic and formal methods to quantum computing and the foundations of artificial intelligence.