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…
从自然语言统计规律出发，揭示神经缩放定律的数学根源，为理解大模型能力增长提供理论基石。
arXiv:2602.07488v3 Announce Type: replace-cross Abstract: Despite the fact that experimental neural scaling laws have substantially guided empirical p…
从逼近到涌现，为深度学习提供统一理论框架，AI研究者必读的硬核前沿。
arXiv:2607.01311v1 Announce Type: new Abstract: Deep learning has outgrown any single mathematical explanation. From Approximation to Emergence develo…
提出“无窥视调优”方法，为大模型后训练提供可证明的泛化界限与鲁棒性保障。
arXiv:2507.01752v4 Announce Type: replace-cross Abstract: Gradient-based optimization is the workhorse of deep learning, offering efficient and scalab…
循环神经网络也能万能逼近连续函数，理论证明补全了深度学习基石。
arXiv:2606.20325v1 Announce Type: new Abstract: Classical approximation theorems ask for a new neural network whenever the target accuracy is improved…
揭示Grokking中权重范数如何通过交叉熵下的logit尺度中介作用控制延迟泛化，为理解神经网络泛化机制提供新视角。
arXiv:2606.18465v1 Announce Type: new Abstract: Grokking, the delayed jump from memorization to generalization, is usually tied to the weight norm: a …
颠覆传统AI范式，提出不使用模型即可实现通用智能的激进新理论，来自UAI 2026的最新研究
arXiv:2602.23242v3 Announce Type: replace Abstract: In general reinforcement learning, all established optimal agents, including AIXI, are model-based…
多项式卷积网络的几何结构深度解析，优化理论新突破被AISTATS 2025收录
arXiv:2410.00722v3 Announce Type: replace Abstract: We study convolutional neural networks with monomial activation functions. Specifically, we prove …
揭秘Transformer缩放定律背后的学习动力学与泛化机制，87页长文深度统一理论框架。
arXiv:2512.22088v3 Announce Type: replace-cross Abstract: The scaling law, a cornerstone of Large Language Model (LLM) development, predicts improveme…
二元脉冲神经网络如何从结构上实现因果建模，这篇论文给出了理论框架与可行路径
arXiv:2604.27007v2 Announce Type: replace Abstract: We provide a causal analysis of Binary Spiking Neural Networks (BSNNs) to explain their behavior. …
利用权重空间对称性简化曲率计算，为优化和泛化理论提供新视角，ICML 2026 前沿成果。
arXiv:2606.00442v1 Announce Type: new Abstract: Many machine learning techniques rely on approximating a loss function's curvature, but this is notori…
用重正化群理论揭示全连接深度神经网络在指数族上的工作机制，为深度学习可解释性提供新视角。
arXiv:2606.00157v1 Announce Type: cross Abstract: We consider establishing the interpretability theory of deep learning through constructing a corresp…
探索神经网络深度与宽度如何影响交互效率与泛化能力，揭示最新定律。
arXiv:2605.27989v1 Announce Type: new Abstract: The guidance of scaling laws has increased the resource demands of modern large language models (LLMs)…
训练后的量子神经网络展现出高斯过程行为，揭示其与经典核方法的深层联系，为量子机器学习理论提供新视角
arXiv:2402.08726v2 Announce Type: replace-cross Abstract: We study quantum neural networks made by parametric one-qubit gates and fixed two-qubit gate…
用Koopman算子理论推导多任务深度学习的泛化边界，为理论分析提供新视角。
arXiv:2512.19199v2 Announce Type: replace-cross Abstract: The paper establishes generalization bounds for multitask deep neural networks using operato…
从经典到前沿，这篇综述系统梳理了神经网络逼近理论的关键成果与最新进展，是理论研究者不可错过的深度参考。
arXiv:2605.21451v1 Announce Type: new Abstract: Universal approximation theorems provide a mathematical explanation for the expressive power of neural…
提出一种严谨且可计算的模型复杂度度量方法，为深度学习理论分析提供新工具
arXiv:2605.21167v1 Announce Type: cross Abstract: An accurate assessment of a model's complexity is crucial for topics such as interpretation, general…
一个arXiv论文链接，助你快速了解神经网络公理化的前沿探索，适合理论研究者跟踪最新进展
arXiv:2605.20534v1 Announce Type: new Abstract: While deep neural networks have achieved remarkable success across a wide range of domains, their unde…
深度探究特征学习如何动态重塑神经网络函数空间的理论前沿成果，59页长文，理论研究者必读。
arXiv:2605.17718v1 Announce Type: cross Abstract: Feature learning is widely regarded as the key mechanism distinguishing neural networks from fixed-k…
深入剖析三层ReLU网络的对称性结构，为深度学习理论提供新视角。
arXiv:2605.18319v1 Announce Type: new Abstract: We develop a framework for analyzing parameter symmetries in deep ReLU networks and obtain a complete …