Learning Optimal Message Representations for Agentic Communication

Large Language Models (LLMs) have demonstrated remarkable capabilities in multi-agent collaborative problem-solving, albeit a gap exists. Existing frameworks predominantly rely on natural language as a primary representation for agentic communication. However, natural language could be ambiguous, biased and ver-bose. Furthermore, recent works have shown that alternative representations can enhance performance in LLMs on certain tasks. However, current approaches lack the intelligence necessary to understand, learn or apply optimal communication representations adaptively. In this paper, we propose to dynamically learn the optimal message representations to enhance agentic performance. We model the optimization problem as an Expanding Markov Decision Process (EMDP) and propose our method named OPTi-MACS. We evaluate our system across benchmark datasets of collaborative problem-solving. The results show significant performance improvements while maintaining efficiency. Our work bridges the gap between rigid communication protocols and open-ended natural language by providing an adaptive framework that learns task-aware structural representations.