HypoLoop:用实验反馈修订权重编辑方案
HypoLoop: Revising Weight-Editing Plans from Scientific Feedback
一群研究者把模型权重直接当成实验对象来改,用实验反馈迭代决定改哪些参数。Burgers 任务上误差降了近一半,还同时测了 DeepSeek 和 Claude。做科研计算或模型微调的人可以看看这套思路。
HypoLoop 是一组原生权重编辑方法,把参数更新决策与科学实验反馈直接关联。方法层面试了固定策略、数值自适应策略和 LLM 引导策略三类方案来生成编辑计划,由数值求解器计算权重补丁,再用内部验证挑选最终部署的预测器,后续观测与验证记录会持续修订可编辑位置和拟合规则。
实验覆盖 22 个科学任务,同时使用 DeepSeek 和 Claude。在 Burgers 任务上,重新分配同等参数预算使误差相对固定位置降低 58.
80% 和 47.72%;
在 DIODE 上,正残差加权在内部交叉验证中将绝对相对误差(AbsRel)降低 13.35%;
跨轮次修订完整计划使最终 AbsRel 相比固定初始计划再降 10.59%。
SpaceNet7 和 Make3D 任务验证了被接受的编辑能改善空间预测。代码已在 GitHub 开源。
HypoLoop: Revising Weight-Editing Plans from Scientific Feedback
Scientific experiments continually reveal prediction errors. Effective updates depend on which parameters change and which errors drive the correction. We introduce HypoLoop, a family of native weight-editing methods that connects these choices to experimental feedback. Fixed, numerically adaptive, and LLM-guided policies produce editing plans; numerical solvers compute weight patches, and internal validation selects the deployed predictor. Successive observations and validation records support revisions to editable locations and fitting rules. Experiments cover 22 scientific tasks and both DeepSeek and Claude. On Burgers, reallocating a matched parameter budget reduces error by 58.80% and 47.72% relative to fixed locations. On DIODE, positive-residual weighting reduces absolute relative error (AbsRel) by 13.35% in internal cross-validation, with the same 33 editable coordinates. Revising the complete plan across rounds lowers final AbsRel by 10.59% relative to keeping the initial plan fixed. SpaceNet7 and Make3D connect accepted edits to improved spatial predictions. These results show how experimental evidence can shape the update problem itself. Code: https://github.com/limengran98/HypoLoop.
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:HypoLoop、DeepSeek、Claude、微调,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:本文原始信息发布自一手权威信源「arXiv: DeepSeek」(Tier-1 一手发布),具备高确定性参考价值。