A Computational Theory for Efficient Mini Agent Evaluation with Causal Guarantees
This work addresses the problem of costly agent evaluation for researchers and practitioners, offering a novel method with significant efficiency gains.
The paper tackles the high cost of experimental evaluation for agents by introducing a computational theory to build evaluation models that accelerate the process, reducing evaluation errors by 24.1% to 99.0% across 12 scenes and cutting evaluation time by 3 to 7 orders of magnitude per subject.
In order to reduce the cost of experimental evaluation for agents, we introduce a computational theory of evaluation for mini agents: build evaluation model to accelerate the evaluation procedures. We prove upper bounds of generalized error and generalized causal effect error of given evaluation models for infinite agents. We also prove efficiency, and consistency to estimated causal effect from deployed agents to evaluation metric by prediction. To learn evaluation models, we propose a meta-learner to handle heterogeneous agents space problem. Comparing with existed evaluation approaches, our (conditional) evaluation model reduced 24.1\% to 99.0\% evaluation errors across 12 scenes, including individual medicine, scientific simulation, social experiment, business activity, and quantum trade. The evaluation time is reduced 3 to 7 order of magnitude per subject comparing with experiments or simulations.