图学习新突破：拓扑到拓扑的LLM智能体，无需特征和任务知识即可高效处理CAD等图结构数据。
arXiv:2512.08964v4 Announce Type: replace Abstract: Graph learning aims to convert data into graph representations, which are fundamental to many prob…
1-bit量化大模型新思路，输出对齐策略再审视，助力低资源设备高效推理
arXiv:2512.21651v3 Announce Type: replace Abstract: Large Language Models (LLMs) deliver strong performance across a wide range of NLP tasks, but thei…
机器遗忘不仅能删除记忆，还能诱导模型产生可控的副作用行为与能力，颠覆了传统遗忘认知。
arXiv:2601.21702v3 Announce Type: replace Abstract: We consider Representation Misdirection (RM), a class of large language model (LLM) unlearning met…
在线矢量量化注意力新机制，巧妙平衡长上下文性能与计算内存效率
arXiv:2602.03922v3 Announce Type: replace Abstract: Standard sequence mixing layers used in language models struggle to balance efficiency and perform…
新算法Mosaic-of-Motifs大幅简化神经网络，参数压缩性能损失极小，揭秘深度学习模型为何易于压缩。
arXiv:2602.14896v2 Announce Type: replace Abstract: Large-scale deep learning models are well-suited for compression. Across a variety of tasks, metho…
一篇研究大语言模型“内省意识”机制的论文，揭示模型如何检测并识别注入的操控向量，行为鲁棒性惊人。
arXiv:2603.21396v4 Announce Type: replace Abstract: Recent work has shown that LLMs can sometimes detect when steering vectors are injected into their…
探究机器学习中流形假设的局限性，揭示模型学习稀疏数据时的几何结构新发现。
arXiv:2605.08464v2 Announce Type: replace Abstract: The manifold hypothesis (MH) is often used to explain how machine learning can overcome the curse …
用元学习+贝叶斯推断搞定复杂系统根因分析，无需完整因果图即可高效定位故障。
arXiv:2605.08786v2 Announce Type: replace Abstract: Root cause analysis (RCA) in complex systems is challenging due to error propagation across multip…
提出力感知神经切线核方法，攻克机器学习力场主动学习在大规模候选池及偏差下的扩展性与鲁棒性难题
arXiv:2605.13788v2 Announce Type: replace Abstract: Active learning for machine-learning interatomic potentials (MLIPs) must address several challenge…
提出有限k分解方法，预测机器学习模型部署时的失败率，提升安全评估可行性。
arXiv:2605.15134v2 Announce Type: replace Abstract: Estimating how often an ML model will fail at deployment scale is central to pre-deployment safety…
首次系统研究结构化Stackelberg博弈，利用上下文信息预测对手类型，助力领导者在线学习最优策略。
arXiv:2504.09006v4 Announce Type: replace-cross Abstract: We initiate the study of structured Stackelberg games, a novel form of strategic interaction…
用结构化模板引导强化学习，让LLM推理训练告别低效自采样，提升策略可迁移性。
arXiv:2505.15692v5 Announce Type: replace-cross Abstract: Reinforcement learning (RL) has emerged as an effective paradigm for enhancing model reasoni…
GAS框架揭示生成式AI中通用性、准确性与简洁性间的根本矛盾，指导组织策略重塑
arXiv:2506.22440v2 Announce Type: replace-cross Abstract: This paper introduces the Generality-Accuracy-Simplicity (GAS) framework to analyze how larg…
固定精度Transformer在描述语言时的简洁性，指数级优于线性时序逻辑和循环神经网络，理论证明其强大表达能力。
arXiv:2510.19315v3 Announce Type: replace-cross Abstract: We study succinctness as a measure of the expressive power of transformers. Succinctness -- …
首次提出确定性Lp子空间核心集的迭代算法，覆盖任意p∈[1,∞)的ε-子空间嵌入
arXiv:2601.00361v3 Announce Type: replace-cross Abstract: We introduce the first iterative algorithm for constructing a $\varepsilon$-coreset that gua…
一篇对LLM毒性问题进行系统性复制研究的最新论文，验证了现有测量与缓解方法、揭示关键发现，值得关注。
arXiv:2605.14087v2 Announce Type: replace-cross Abstract: Large Language Models (LLMs) trained on web-scale corpora inherently absorb toxic patterns f…
新研究揭示共形预测池化校准隐藏组间异质性，推导守恒律指出公平性代价
arXiv:2605.14260v2 Announce Type: replace-cross Abstract: Conformal prediction is often calibrated with a single pooled threshold, but this can hide c…
106本小说16种语言，130k段落揭示大模型文学翻译流畅性与忠实度间的微妙平衡
arXiv:2605.15282v1 Announce Type: new Abstract: Literary translation requires balancing target-language fluency with faithfulness to the source. Recen…
加纳法律教育迎来AI助手Eskwai，探索生成式AI在发展中国家教育领域的应用潜力。
arXiv:2605.15380v1 Announce Type: new Abstract: Recent advances in generative AI have shown their potential to be leveraged for legal education. Yet, …
大型语言模型个性化新方案：Capability Conditioned Scaffolding，解决专业领域依赖AI推理导致的能力漂移风险。
arXiv:2605.15404v1 Announce Type: new Abstract: Large language model personalization typically adapts outputs to user preferences and style but does n…