视觉Transformer部署提速新思路，用脆弱性引导混合精度量化，兼顾压缩率与精度的双赢方案。
arXiv:2607.28589v1 Announce Type: cross Abstract: Post-training quantization (PTQ) has emerged as an effective solution for deploying Vision Transform…
用在线线性规划算法优化LLM服务的多目标路由，兼顾效率与资源均衡
arXiv:2607.03948v1 Announce Type: new Abstract: We study the online routing problem in large language model serving, where requests arrive sequentiall…
交叉验证与稀疏回归结合，如何显著降低计算成本？这篇论文给出了高效算法。
arXiv:2306.14851v5 Announce Type: replace-cross Abstract: Given a high-dimensional covariate matrix and a response vector, ridge-regularized sparse li…
拆解RL策略梯度权重，FADE方法防止性能衰减，揭秘增益消失的根源。
arXiv:2607.01490v1 Announce Type: cross Abstract: Reinforcement learning post-training dramatically improves LLM reasoning, but suffers from training …
异构场景下分布式学习的新方案，把节点差异化为加速信号，值得关注。
arXiv:2505.09854v3 Announce Type: replace Abstract: As end-user device capability increases and demand for intelligent services at the Internet's edge…
公平联邦学习新突破：用方差正则化平衡全局精度与个体公平，理论扎实，实验有说服力。
arXiv:2508.12042v2 Announce Type: replace Abstract: Federated learning (FL) allows collaborative training of machine learning models across multiple p…
LLM增强的A*算法突破几何限制，巧用语言模型引导非几何网络图的最优路径搜索
arXiv:2606.23136v1 Announce Type: cross Abstract: Finding the shortest path in non-geometric network graphs, where edge weights encode arbitrary metri…
利用相似性结构提升多臂赌博机算法效率，最新arXiv论文探索更优的探索-利用权衡策略。
arXiv:2606.23414v1 Announce Type: new Abstract: In many online learning and bandit problems, the actions we consider possess inherent similarities--fo…
Leetcode 169多数元素题解，从哈希表到Boyer-Moore投票算法的两种思路对比
Leetcode 169 asks us to identify the majority element in a given array. The majority element is defined as the element that appears more than n / 2 ti…
Q-learning新理论突破，细粒度依赖遗憾界优化算法性能。
arXiv:2510.06647v2 Announce Type: replace-cross Abstract: We study fine-grained gap-dependent regret bounds for model-free reinforcement learning in e…
重新审视LLM RL中的散度正则化，提出改进方案，提升模型对齐效率
arXiv:2606.09821v1 Announce Type: new Abstract: Reinforcement learning (RL) has become a key component of post-training large language models (LLMs). …
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 …
用双指针模式替代O(n²)嵌套循环，通过经典类比和实战技巧大幅提升代码效率
A hands-on guide to one of the most powerful algorithmic patterns you'll use every week — with real code in Java, Python, and C. ✍️ Written for develo…
扩散LLM推理加速新突破：Fréchet Profile Decoding让推理更高效
arXiv:2606.02955v1 Announce Type: cross Abstract: Diffusion large language models promise parallel token generation, yet inference remains bottlenecke…
揭露LLM推荐系统过度依赖语义的缺陷，提出保留协同频率成分的新思路
arXiv:2508.10312v2 Announce Type: replace Abstract: Recommender systems in concert with Large Language Models (LLMs) present promising avenues for gen…
Nature报道：重新设计算法让社交信息流变得“无毒”，平衡内容质量与用户体验
Nature, Published online: 27 May 2026; doi:10.1038/d41586-026-01636-z A large-scale experiment tested the impact of changing the Bluesky feed-ranking …
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…
强化学习新突破：XRPO通过定向探索与利用，显著提升GRPO在LLM推理任务中的性能与稳定性
arXiv:2510.06672v3 Announce Type: replace Abstract: Reinforcement learning algorithms such as GRPO have driven recent advances in large language model…
视觉-only BEV感知新框架Fast-BEV++，在精度与部署效率间找到平衡，加速自动驾驶落地。
arXiv:2512.08237v3 Announce Type: replace Abstract: The advancement of vision-only Bird's-Eye-View (BEV) perception, a core paradigm for cost-effectiv…
将次线性图算法的先验知识理论引入LLM测试时优化，开辟AI效率提升新路径。
arXiv:2510.16609v3 Announce Type: replace Abstract: Test-time augmentation, such as Retrieval-Augmented Generation (RAG) or tool use, critically depen…