IF-MAP: Influence-Function-Inspired Agent Prioritization in Multi-Agent Systems

Key points

• Improving every agent in an LLM-based multi-agent system can be prohibitively costly, making it important to identify which agent should be prioritized to achieve the greatest system-level gain with limited resources.1
• For each agent, IF-MAP measures end-to-end loss changes under controlled semantic prompt perturbations, fits a regularized local response model in a low-dimensional prompt-embedding subspace, and uses the estimated response norm as a priority score.1
• We evaluate IF-MAP across six multi-agent workflows spanning code generation, question answering, and mathematical reasoning, using the same priorities for prompt optimization and model upgrades.1
• Despite its probing overhead, IF-MAP reduces Qwen3-8B and DeepSeek-V3.2 token use by 74.2% and 35.2%, respectively, relative to optimizing every agent, and reduces target-model tokens by 63.6-64.9% relative to upgrading every agent.1

Sources (1)

• [1]IF-MAP: Influence-Function-Inspired Agent Prioritization in Multi-Agent Systems
  arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 17, 02:36 AM

  We therefore introduce IF-MAP, an influence-function-inspired method for identifying agents with high potential for system-level improvement.

  Improving every agent in an LLM-based multi-agent system can be prohibitively costly, making it important to identify which agent should be prioritized to achieve the greatest system-level gain with limited resources.

Extractive summary: sentences quoted from the sources.