一份用信赖域框架重新审视自适应矩估计的数学研究，适合想深挖优化器收敛原理的读者。
arXiv:2608.04026v1 Announce Type: cross Abstract: In this paper, we develop a trust-region framework for understanding the behavior of adaptive moment…
被ICML接收的自适应梯度裁剪方法，有效提升LLM预训练稳定性，AI训练优化的新突破
arXiv:2502.11034v3 Announce Type: replace Abstract: Loss spikes remain a persistent obstacle in large-scale language model pretraining. While previous…
贝叶斯后验引导技能演化，让LLM Agent在不调整权重下高效适配外部推理条件，开辟Agent自适应新路径。
arXiv:2606.08348v1 Announce Type: new Abstract: LLM agents increasingly rely on external inference conditions: prompts, tools, memory, SOPs, skills, a…
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…
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…