1
Successor-Generator Planning with LLM-generated Heuristics
用大模型生成启发式函数，为经典规划问题开辟新思路。
arXiv:2501.18784v5 Announce Type: replace Abstract: Heuristics are a central component of deterministic planning, particularly in domain-independent s…
2
Beyond Heuristic Tuning: Power-Calibrated LLM Watermarking
ICML 2026接收：为LLM水印引入功率校准，告别传统启发式调优。
arXiv:2607.05694v1 Announce Type: cross Abstract: Logit-based watermarking is a widely used mechanism for identifying LLM generated content, yet its e…
3
Graph Neural Networks are Heuristics
颠覆认知：GNN本质是高级启发式算法而非特征学习器，理论推导+实验验证给出新视角
arXiv:2601.13465v4 Announce Type: replace Abstract: Graph neural networks are usually treated as auxiliaries for combinatorial optimization: they imit…
4
The Model Says Walk: How Surface Heuristics Override Implicit Constraints in LLM Reasoning
揭示大语言模型推理时被表面启发式误导而忽略隐含约束的机制，对理解AI的认知偏差有重要启发。
arXiv:2603.29025v3 Announce Type: replace-cross Abstract: Large language models fail when a salient surface cue conflicts with an unstated feasibility…
5
Discovering heuristics in a complex SAT solver with large language models
大语言模型不再只是聊天，还能帮助复杂SAT求解器自动发掘高效启发式策略，开启AI辅助算法设计新范式。
arXiv:2507.22876v2 Announce Type: replace Abstract: The Satisfiability problem (SAT) is fundamental in computational complexity theory and has a wide …
6
Slopper GitHub Action: Fighting AI Slop Contributions on Open Source Projects
用6个启发式信号检测AI生成的垃圾PR，无需API密钥，开源维护神器。
Article URL: https://github.com/malvads/Slopper Comments URL: https://news.ycombinator.com/item?id=48426225 Points: 4 # Comments: 0
7
Learning Admissible Heuristics via Cost Partitioning
最新研究提出通过成本划分学习可容许启发式，解决最优规划中过度估计难题，为自动化推理提供新思路。
arXiv:2606.04597v1 Announce Type: new Abstract: Admissible heuristics are essential for optimal planning, yet learning them remains challenging due to…
8
LLM-Evolved Domain-Independent Heuristics for Symbolic AI Planning
将LLM与符号AI规划结合，自动进化出领域无关的启发式函数，突破传统规划效率瓶颈。
arXiv:2605.29649v1 Announce Type: new Abstract: Heuristic search is the dominant paradigm in symbolic AI planning, and the strongest heuristics are th…
9
Hierarchical Task Network Planning with LLM-Generated Heuristics
LLM不仅能对话，还能当规划师：用生成启发式破解HTN规划难题，效率与鲁棒性双提升
arXiv:2605.07707v2 Announce Type: replace Abstract: HTN planning is a variation of classical planning where, instead of searching for a linear sequenc…
10
ReVEL: Multi-Turn Reflective LLM-Guided Heuristic Evolution via Structured Performance Feedback
新论文提出多轮反思式LLM框架，通过结构化性能反馈引导启发式进化，提升复杂问题求解效率。
arXiv:2604.04940v2 Announce Type: replace Abstract: Designing effective heuristics for NP-hard combinatorial optimization problems remains challenging…
11
Interestingness as an Inductive Heuristic for Future Compression Progress
Schmidhuber新作：将“趣味性”形式化为未来压缩进步的归纳启发式，用Kolmogorov复杂度预判数据潜力，直指递归自我改进AI的核心瓶颈。
arXiv:2605.14831v1 Announce Type: new Abstract: One of the bottlenecks on the way towards recursively self-improving systems is the challenge of inter…
12
Heuristic Pathologies and Further Variance Reduction via Uncertainty Propagation in the AIVAT Family of Techniques
论文揭示AIVAT方差缩减技术中的启发式病态，并提出不确定性传播进一步降低方差，提升多智能体性能评估精度。
arXiv:2605.14261v1 Announce Type: new Abstract: How should an agent's performance in a multiagent environment be evaluated when there is a limited sam…
13
Do Biological Structural Guarantees Earn Their Complexity?
生物启发式AI框架声称结构保证更可靠，这篇论文用三个深度基准实证检验其是否优于朴素替代方案。
arXiv:2605.15225v1 Announce Type: cross Abstract: Biologically-inspired AI agent frameworks claim reliability benefits through structural guarantees a…
14
Petri Net Induced Heuristic Search for Resource Constrained Scheduling
将资源约束项目调度问题建模为 Petri Net 可达图搜索，并用 A* 与关键路径启发式高效求解。
arXiv:2605.15983v1 Announce Type: new Abstract: We formulate the Resource-Constrained Project Scheduling Problem (RCPSP) as optimal search over the re…