EnvHarness：为智能体学习唤醒静态环境
EnvHarness: Awakening Static Worlds for Agent Learning
浏览论文内容
中文总结 AI 辅助
EnvHarness是包裹静态环境的可编程插件层，结合EnvRigger自动合成组件，在多领域基准测试中提升智能体学习性能，减少执行步骤并支持策略与环境协同进化。
中文摘要 AI 辅助
大语言模型（LLM）智能体通过与环境交互进行学习，但这些环境是手工构建且静态的：无法感知智能体的弱点，且会随智能体能力提升而快速失效。尽管近期的环境生成方法试图解决该问题，但它们需要特定领域的流程、依赖昂贵或不可靠的验证器，且仍生成静态环境。为减轻从头重建环境的工程负担，我们提出Environment Harness（EnvHarness），这是一个可编程的插件组件层，可包裹静态环境以重塑其行为，且无需修改底层逻辑。通过标准接口运行，EnvHarness可适用于不同领域，同时确保每个被重塑的环境保留其原始验证器。为实现该过程的自动化，我们引入EnvRigger，它将目标策略视为黑盒，观察其执行轨迹以合成针对已诊断缺陷的EnvHarness组件，并通过新的滚动部署对其进行验证。在四个领域的五个基准测试中，EnvHarness的性能优于原始环境和特定领域的环境生成流程，在未见过的实例上实现了高达9.0分的提升，同时执行步骤减少了9.8%。此外，EnvHarness为强化学习提供了更优的优化信号，支持策略与其环境的持续、针对性协同进化。
英文摘要
LLM agents learn by interacting with environments, yet these environments are hand-built and static: blind to an agent's weaknesses, and quickly left behind as it improves. While recent environment generation methods attempt to address this, they require domain-specific pipelines, rely on expensive or unreliable verifiers, and still produce static environments. To alleviate the engineering burden of rebuilding environments from scratch, we propose Environment Harness (EnvHarness), a programmable layer of plug-in components that wraps a static environment to reshape its behavior without modifying the underlying logic. Operating through standard interfaces, EnvHarness applies across diverse domains while ensuring every reshaped environment retains its original verifier. To automate this process, we introduce EnvRigger, which treats the target policy as a black box, observing its execution trajectories to synthesize EnvHarness components targeting diagnosed flaws, and validating them via fresh rollouts. Across five benchmarks in four domains, EnvHarness outperforms both original environments and domain-specific environment generation pipelines, achieving up to a 9.0-point improvement on held-out instances with 9.8% fewer execution steps. Furthermore, EnvHarness provides a superior optimization signal for reinforcement learning, enabling continuous, targeted co-evolution of the policy and its environment.
