Bayesian control for coding agents
For developers of coding agents, this work provides a principled, uncertainty-aware orchestration method that improves cost-effectiveness and interpretability over fixed rules.
The paper formulates coding agent orchestration as cost-sensitive sequential hypothesis testing, using a Bayesian controller that dynamically decides when to gather evidence, refine, verify, or stop. Across six generators and nine benchmarks, Bayesian control is most valuable when verification is costly and critics are informative but imperfect, and its belief state yields an interpretable correctness score that outperforms baselines for uncertainty quantification.
Modern coding agents pair LLM generators with various tools, including cheap diagnostics and expensive verifiers. The tool-use decisions are typically governed by orchestrators that often use fixed rules and ignore uncertainty. We formulate orchestration as cost-sensitive sequential hypothesis testing: a Bayesian controller maintains a belief over candidate correctness and dynamically decides whether to gather more evidence, refine the candidate, verify it, or stop. Across six generators and nine coding benchmarks, Bayesian control proves to be most valuable when verification is costly and critics are informative but imperfect. Beyond control, the belief state yields an interpretable correctness score that outperforms token-probability and raw tool-success baselines for uncertainty quantification.