1
Learning to Configure Agentic AI Systems
提出学习式配置方法，替代手动调优，提升LLM智能体系统的自适应能力
arXiv:2602.11574v3 Announce Type: replace Abstract: Configuring LLM-based agent systems involves choosing workflows, tools, token budgets, and prompts…
2
微软 XBOX 上线自适应摇杆帽结构，新增球门柱造型
IT之家 5 月 22 日消息，在全球无障碍宣传日（每年 5 月的第三个星期四，2026 年为 5 月 21 日），微软 XBOX 面向玩家，发布多项无障碍更新，聚焦硬件适配、信息获取与游戏体验 3 个层面。 本次更新最值得关注的项目，是自适应摇杆帽计划，微软已在 XBOX Design Lab 上…
3
STAR: Semantic-Tuned and Tail-Adaptive Retriever for Graph-Augmented Generation
图增强生成新方案：语义调优与尾部自适应检索，提升复杂查询效果。
arXiv:2605.18765v1 Announce Type: cross Abstract: To augment Large Language Models (LLMs) for multi-hop question answering, a mainstream solution with…
4
Adaptive Probe-based Steering for Robust LLM Jailbreaking
针对LLM越狱新方法，采用自适应探针引导，克服了传统对比引导的偏差和手动调参局限，提升鲁棒性与有效性。
arXiv:2605.20286v1 Announce Type: cross Abstract: Recent work has demonstrated the potential of contrastive steering for jailbreaking Large Language M…
5
From SGD to Muon: Adaptive Optimization via Schatten-p Norms
深入解读从SGD到Muon的优化器演进，以Schatten-p范数统一矩阵几何约束，为AI研究者提供理论新视角
arXiv:2605.19781v1 Announce Type: new Abstract: Modern optimizers, like Muon, impose matrix-wise geometry constraints on their updates. These matrix-w…
6
Sequential Consensus for Multi-Agent LLM Debates: A Wald-SPRT compute governor with calibration-based failure detection
用Wald的序贯概率比检验动态调控LLM辩论轮数，显著提升推理效率与资源利用率
arXiv:2605.19193v1 Announce Type: new Abstract: Multi-agent LLM debate improves factuality and reasoning, but most recipes pick a fixed round count, o…
7
Skills on the Fly: Test-Time Adaptive Skill Synthesis for LLM Agents
LLM智能体测试时动态合成技能，无需预训练即可灵活适应新任务
arXiv:2605.16986v1 Announce Type: new Abstract: LLM agents benefit from reusable skills, yet test-time tasks often require guidance more specific than…
8
DashAttention: Differentiable and Adaptive Sparse Hierarchical Attention
提出可微分自适应稀疏层次注意力机制，显著提升长序列建模效率与计算可扩展性
arXiv:2605.18753v1 Announce Type: cross Abstract: Current hierarchical attention methods, such as NSA and InfLLMv2, select the top-k relevant key-valu…
9
LEAP: Learnable End-to-End Adaptive Pruning of Large Language Models
提出LEAP可学习端到端自适应剪枝方法，在保持大语言模型性能的同时实现高效压缩
arXiv:2605.17289v1 Announce Type: new Abstract: Unstructured sparsity is now natively accelerated by recent GPU kernels and dataflow hardware, shiftin…
10
HPC-LLM: Practical Domain Adaptation and Retrieval-Augmented Generation for HPC Support
如何让大模型理解高性能计算？这篇论文用领域自适应和RAG打造HPC专属LLM，实用又前沿。
arXiv:2605.16347v1 Announce Type: new Abstract: Modern scientific research increasingly depends on High-Performance Computing (HPC) infrastructures, y…
11
Dual-Dimensional Consistency: Balancing Budget and Quality in Adaptive Inference-Time Scaling
LLM推理时缩放的新方法：双维度一致性，有效平衡采样预算与推理质量，提升效率。
arXiv:2605.15100v1 Announce Type: new Abstract: Large Language Models (LLMs) have demonstrated remarkable abilities in reasoning. However, maximizing …
12
Stage-adaptive Token Selection for Efficient Omni-modal LLMs
阶段自适应token选择策略，显著提升全模态大语言模型推理效率，突破多任务性能瓶颈。
arXiv:2605.20035v1 Announce Type: new Abstract: Omni-modal large language models (om-LLMs) achieve unified audio-visual understanding by encoding vide…
13
Aperion Shield: local guardrail that blocks destructive AI coding agent ops
开源本地安全护栏，阻挡AI编码代理的破坏性操作，v0.7新增45+自适应规则与更多测试用例。
Article URL: https://github.com/AperionAI/shield Comments URL: https://news.ycombinator.com/item?id=48207471 Points: 2 # Comments: 0
14
CoLLM: Continuous Adaptation for SLO-Aware LLM Serving on Shared GPU Clusters
面向共享GPU集群，提出连续自适应方法优化大模型服务SLO，降低延迟与成本
arXiv:2604.16400v2 Announce Type: replace-cross Abstract: As Large Language Models (LLMs) are increasingly adopted in edge intelligence to power domai…
15
RAP: Runtime Adaptive Pruning for LLM Inference
提出运行时自适应剪枝方法，让LLM推理内存动态调整，效率大增
arXiv:2505.17138v5 Announce Type: replace Abstract: Large language models (LLMs) excel at language understanding and generation, but their enormous co…
16
ANNEAL: Adapting LLM Agents via Governed Symbolic Patch Learning
提出ANNEAL框架，用符号补丁学习让LLM智能体高效自适应新环境，无需微调大模型。
arXiv:2605.16309v1 Announce Type: cross Abstract: LLM-based agents can recover from individual execution errors, yet they repeatedly fail on the same …
17
AMO: Adaptive Muon Orthogonalization
提出自适应Muon正交化方法，有望优化深度学习训练过程。
arXiv:2605.17806v1 Announce Type: new Abstract: Muon has recently emerged as a competitive alternative to AdamW for large-scale pre-training, with ort…
18
D$^2$Evo: Dual Difficulty-Aware Self-Evolution for Data-Efficient Reinforcement Learning
提出双难度感知自进化方法，解决强化学习训练数据稀缺与动态难度转移的挑战。
arXiv:2605.17037v1 Announce Type: new Abstract: Reinforcement learning (RL) has demonstrated potential for enhancing reasoning in large language model…
19
From Ranking to Reasoning: Explainable Web API Recommendation via Semantic Reasoning
颠覆传统排序推荐，用语义推理让Web API推荐过程可解释、自适应组合复杂度
arXiv:2511.05820v2 Announce Type: replace-cross Abstract: The rapid growth of Web APIs has made automated Web API recommendation essential for efficie…
20
Personalized Digital Health Modeling with Adaptive Support Users
从人口预训练到自适应支持，新方法解决数字健康中数据稀疏下的个性化建模难题
arXiv:2605.02004v2 Announce Type: replace Abstract: Personalized models are essential in digital health because individuals exhibit substantial physio…