发表机构
Impossible Research; UC Berkeley; Carnegie Mellon University（Impossible Research; 加州大学伯克利分校; 卡内基梅隆大学）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
Schema通过交互式程序归纳，将智能体对环境理解编码为可执行程序，显著提升未知环境任务完成率，ARC-AGI-3从58.7%升至99.2%。
AI 中文摘要
在规则未知的陌生环境中学习完成任务，对LLM智能体而言仍是一项关键挑战。当前的LLM智能体通常以散文形式记录其发现，这可能无法对环境运作方式提供紧凑且明确的说明。受科学家将观察组织成可检验、可预测理论的启发，我们引入了Schema，一种通过交互式程序归纳来组织学习与行动的智能体框架。LLM智能体决定调查什么以及如何行动，将其对环境不断演化的理解表达为可执行程序。该框架由一个持久化程序工作区及一组小型接口组成，用于对照交互历史检查这些程序、在程序内部进行规划，并在逐步验证下执行计划。Schema将ARC-AGI-3 RHAE从58.7%提升至99.2%（使用相同基础模型），解决了DiG-bench公开游戏中的100%，并在MazeBench上达到前50名人类玩家的中位性能。广泛分析表明Schema在未知机制发现中的有效性，消融实验确认了每个组件的贡献。
英文摘要
Learning to complete tasks in unfamiliar environments with unknown rules remains a key challenge for LLM agents. Current LLM agents often record their discoveries in prose, which may not provide a compact, explicit account of how the environment works. Inspired by how scientists organize observations into testable, predictive theories, we introduce Schema, an agent harness that organizes learning and action through interactive program induction. The LLM agent decides what to investigate and how to act, expressing its evolving understanding of the environment as executable programs. The harness consists of a persistent program workspace and a small set of interfaces for checking these programs against the interaction history, planning within them, and executing plans under step-by-step verification. Schema raises ARC-AGI-3 RHAE from 58.7% to 99.2% with the same base model, solves 100% of the public DiG-bench games, and reaches the median performance of the top-50 human players on MazeBench. Extensive analysis shows the effectiveness of Schema in unknown mechanism discovery, and ablations confirm the contribution of each component.
CommentsProject Website: https://schema-harness.github.io/
