CoReMem：为连续编码任务设计代价内化的记忆框架
CoReMem: Cost-Internalizing Memory for Successive Coding Tasks with Prospective Valuation
arXiv上这篇论文解决了编码智能体记忆该存什么、什么时候用的问题。它把读写记忆的成本算进训练目标，在ChainSWE基准上让27B模型解决率从21.0%提到25.7%，还少走17%的步数。做智能体记忆或SWE agent的团队值得对比看看。
一项arXiv论文提出Coding-Repository Memory（CoReMem），针对编码智能体在演进代码库上连续执行任务时如何选择记忆的问题，将写入与读取记忆的代价直接纳入优化目标。框架通过prospective segment valuation比较下一任务在有与无各条轨迹片段时的运行结果，据此构建以锚点（代码库位置和可执行检查）为核心的紧凑记忆。
Actor适配环节在含与不含记忆的rollout组内估计on-policy记忆价值，奖励设计优先提升解决率，并对成功运行的步数施加惩罚；Provenance-gated value credit则把额外价值记到实际使用记忆锚点的步骤上，其余步骤仍保留结果奖励。在ChainSWE基准上，27B actor的解决率从21.0%提升到25.7%，步数减少17%，含记忆模块调用的总cache-aware token数减少11%。
对比常见的无差别记忆写入方案，CoReMem明确衡量每条记忆对未知下一任务的边际价值，适合研究智能体记忆机制和长程编码工作流的开发者参考。
CoReMem: Cost-Internalizing Memory for Successive Coding Tasks with Prospective Valuation
Coding agents tackle successive tasks on an evolving repository, carrying forward edits and writing memory before the next task is known. Selecting memory for unknown next tasks requires balancing future usefulness against the costs of writing and reading it. The actor must also decide when acting on available memory saves effort as the task unfolds. We introduce Coding-Repository Memory (CoReMem), a cost-internalizing framework that guides memory design and actor adaptation through measured memory value on the next task. Prospective segment valuation compares runs of the next task with and without individual trajectory segments from the same inherited repository state, guiding a compact memory built around anchors: repository locations and executable checks. Actor adaptation estimates on-policy memory value within rollout groups with and without memory, using a reward that prioritizes resolution and penalizes step count in successful runs. Provenance-gated value credit assigns the additional value term to steps identified as using memory-supplied anchors, while every step retains outcome credit. On ChainSWE, CoReMem raises the 27B actor's resolve rate from 21.0% to 25.7%, with 17% fewer steps and 11% fewer total cache-aware tokens, including memory-module calls, relative to the untrained actor model without memory.
本条情报归属 AITOP「论文」核心追踪栏目。在跨区域产业观察坐标中，该动态直接关联「AI 开发者工具链、Coding Agent 与工程化实践」研究主线，反映出当前技术演进与商业落地的最新风向。涉及核心实体与技术议题包括：CoReMem、编码智能体、ChainSWE、记忆机制，读者可通过下方标签进入对应专题页查阅历史脉络与关联对比。
信源核验说明：本文原始线索采编自「arXiv cs.AI」（@Peiwen Li, Harshit Verma, Sophia Zhou, Haiwen Wang, Ngoc Bui, Heng Chang, Fei Fei, David van Dijk, Rex Ying），经 AITOP 多维评分管线结构化提炼。