Tag
The paper proposes CAP (Candidate Acquisition Policy), a method that learns a feedback-conditioned query policy from sparse records of related black-box optimization tasks, achieving the best final-budget results on four synthetic families and two engineering benchmarks (BOPTEST, SU2).
The paper introduces AutoHebb, a class of self-modifying neural networks using a three-factor Hebbian learning rule gated by an autonomously generated neuromodulatory signal, which achieves zero-shot inference and continual learning while outperforming competing models.
The authors announce Gear, an open-source, meta-learning-inspired framework described as "a scikit-learn for RSI," which separates search decisions from execution and evaluation so methods like GEPA, RRSI, and Harness² are easier to implement and compare. A working GEPA-style variant with shared failure diagnosis and staged evaluation showed pass-rate gains on AutomationBench and Terminal-Bench, plus a ~95% lower per-task API cost for GPT-5.6 Luna versus Astra using the same optimized harness.
CHORD提出了一种跨模型幻觉检测（CMHD）框架，将LLM隐藏状态表示为层级-词元-特征关系图，通过关系感知注意力与元学习获得可迁移到未见目标模型的检测信号，在跨模型泛化上优于现有检测方法。
提出 SHINE，一种从稀疏历史搜索日志与零成本代理出发的跨任务神经架构搜索策略，通过组合各源任务的冠军架构并联合训练成对排序器，显著缓解单一知识源在目标任务上的严重失败，同时保持有竞争力的平均性能。
RSI-Opt 通过元智能体在多次训练运行间累积并进化优化器策略（提示、流程、代码库与决策节奏），显著提升 LLM 优化器智能体在模型训练与科学机器学习任务上的性能，并支持跨模型策略迁移。
Skill∞ 介绍了一种递归自我改进框架，让元智能体通过可复用的元技能同时优化任务技能与优化流程本身，在 83 个 SkillsBench 任务上达到 60.7% 的通过率，优于人工整理技能和 SkillEvolver。
Meta-iLaD 提出一种通过跨动力学环境元学习实现可识别潜在动力学的新框架，利用少量样本的条件先验快速提取环境变量，并同时对潜在状态、环境变量与动力学函数建立可识别性，从而提升对未见环境的泛化与预测能力。
该论文提出垂直切片Wasserstein距离(VSW)，一种用于比较meta-measures（分布的分布）的可扩展度量，证明其为有效度量并具有维度无关的统计收敛保证，在效率和精度上优于现有方法。
该论文提出开放世界AI需要三项核心要求：推理时学习、丰富表示和预测性解纠缠，并在最大880亿参数的语言模型上验证了满足这些要求的模型能从相关任务数据中持续提升，而其他模型则会退化。
Google 的 Harness² 论文研究智能体在开放世界环境中对自身 Harness 的自进化，采用双层递归机制：无需标准答案即可通过行为对比筛选有效修改，并将修改过程本身总结为经验以指导未来任务，在多个基准上显著提升 GLM-5 和 Gemini 3.5 Flash 的性能。
推文介绍了一个关于递归自我改进(RSI)的精选研究合集，涵盖从STaR、Self-Rewarding LMs到Promptbreeder和Automated Design of Agentic Systems等关键论文，梳理了RSI从自生成训练到自我修改代码代理的演进脉络。
论文提出 MotherTree，一个通过合成先验数据进行元学习的表格 Transformer，可在单次前向传播中输出可解释的硬决策树，在小样本任务上表现优于从零训练的梯度方法，并可作为强初始化器进一步提升决策树学习。
该论文提出一种预处理感知的基准测试框架，在M4数据集的29,000条时间序列上评估11个预测模型与16种可逆预处理管线，发现为每条序列优化预处理可带来约27%至87%的性能提升，并揭示现有基准存在结构性预处理偏差。
This paper introduces the Prior-Fitted Language Model (PFLM), a 300M-parameter byte-level transformer pretrained exclusively on samples from a synthetic structural causal model, enabling it to learn languages in context with frozen weights despite never seeing real text. Results show near-competitivenatural-language modeling, arithmetic, and compression capabilities below gzip/PPMd, suggesting meta-learning of language rather than memorization.
Meta 论文提出 RankEvolve，通过可执行操作协议（EOP）将研究阶段编译为受运行时强制执行的状态机，并让 Claude Code 与 Codex 作为节点互相审查修复，在预算持平下将执行准确率从 45.8% 提升到 62.5%，且在 MovieLens-20M 上改进推荐模型的 NDCG@10 达 4.48%。
CMU paper introduces "harness learning," where an RL-trained proposer model revises an agent's executable harness from execution feedback without changing solver weights, achieving generalizable test-time adaptation across unseen tasks with a 4B proposer outperforming its 35B teacher.
This paper proposes a meta-learning framework for online convex optimization with indicator switching costs, achieving minimax-optimal dynamic regret bounds without prior knowledge of comparator variations.
This paper proposes PROBE, a perturbed gradient algorithm for bilevel optimization with nonconvex lower levels, using second-order stationarity to achieve finite-time convergence and outperform state-of-the-art methods in experiments on LLM-based tasks and meta-learning.
This paper introduces Meta-Skill, a method letting a frozen Builder model learn reusable principles from Target execution feedback to construct better agent harnesses for unseen tasks, improving performance by 8.95 points on Harness-Bench and NewtonBench. It suggests a path toward system-level self-improvement by learning to build better environments rather than changing model weights.