发表机构
Indian Institute of Technology Kharagpur; Google Research（印度理工学院卡拉格普尔分校; 谷歌研究院）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
本研究提出ORDDAR推理框架，通过局部认知状态转换检测、失真定位与针对性修复，提升了多类推理任务的质量、恢复能力与可解释性。
AI 中文摘要
AI智能体越来越多地执行长期推理、规划、工具使用、记忆整合与自主决策，但错误的中间状态会传播并导致决策不一致、输出不可靠。现有推理方法主要依赖迭代规划、自我反思、增强记忆或验证，却很少定位并选择性修复错误推理。我们提出ORDDAR（Observation-Driven Reasoning for Distortion-Resilient Decision, Action, and Cognitive Recovery），这一推理框架将推理建模为认知状态转换，检测局部失真，从过往经验中检索相关推理，且仅修复受影响的状态。因此ORDDAR在局部推理转换层面执行恢复，而非重新生成完整轨迹。在数学、常识、多跳及临床推理基准上的实验表明，与多个被评估的推理基线相比，ORDDAR的推理质量、恢复能力与可解释性均有所提升。
英文摘要
AI agents increasingly perform long-term reasoning, planning, tool use, memory integration, and autonomous decision making, yet erroneous intermediate states can propagate and cause inconsistent decisions and unreliable outputs. Existing reasoning approaches mainly rely on iterative planning, self-reflection, augmented memory, or verification, but rarely localize and selectively repair faulty reasoning. We present ORDDAR (Observation-Driven Reasoning for Distortion-Resilient Decision, Action, and Cognitive Recovery), a reasoning framework that models reasoning as cognitive state transitions, detects localized distortions, retrieves related reasoning from prior experiences, and repairs only the affected states. ORDDAR therefore performs recovery at the local reasoning-transition level rather than regenerating the complete trajectory. Experiments across mathematical, commonsense, multi-hop, and clinical reasoning benchmarks demonstrate improved reasoning quality, recovery ability, and interpretability over multiple evaluated reasoning baselines.
