ADIAS: Automated Design of Interactive Agentic Systems

ADIAS：交互式智能体系统的自动化设计

Abstract: Automated agent design improves agent harnesses through iterative revision, evaluation, and feedback summarization. Existing methods are largely candidate-centric: cross-round experience is organized around candidate agents, which leaves the repair progress implicit. This causes inefficient repair targeting, slow consolidation of partial progress, and propagation of ineffective interventions across rounds.

摘要： 自动化智能体设计通过迭代修订、评估和反馈总结来改进智能体框架。现有的方法大多以候选者为中心：跨轮次的经验围绕候选智能体进行组织，这使得修复进度处于隐性状态。这导致了修复目标定位效率低下、部分进展的整合缓慢，以及无效干预在各轮次间的传播。

Therefore, we formulate issue-centric agent optimization, in which repair progress is carried forward as an explicit persistent issue state to guide optimization, rather than re-derived from candidate history in each round. We instantiate the formulation in ADIAS, a framework for automated full-code agent design with two mechanisms. A persistent issue state maintains stable issue identities, lifecycle status, supporting evidence, and intervention-outcome histories. Issue-guided optimization uses this state to jointly propose repair targets and revision directions for subsequent focused full-code modification.

因此，我们提出了以问题为中心的智能体优化方法，将修复进度作为显式的持久化问题状态进行传递，以指导优化，而不是在每一轮中从候选历史中重新推导。我们在 ADIAS 中实例化了这一方案，这是一个包含两种机制的自动化全代码智能体设计框架。持久化问题状态维护了稳定的问题标识、生命周期状态、支持证据以及干预结果历史。问题引导优化利用该状态，为后续的针对性全代码修改联合提出修复目标和修订方向。

Across five interactive benchmarks, ADIAS outperforms the strongest baseline by 25.2% on average and achieves consistent gains across four backbone models. Controlled ablations further show that removing persistent issue state or replacing issue-centric revision with candidate-centric policies leads to performance drops of up to 40.7%.

在五个交互式基准测试中，ADIAS 的表现平均比最强的基准模型高出 25.2%，并在四个骨干模型上均取得了持续的性能提升。对照消融实验进一步表明，移除持久化问题状态或将以问题为中心的修订替换为以候选者为中心的策略，会导致性能下降高达 40.7%。