惯性聚合 高效追踪和阅读你感兴趣的博客、新闻、科技资讯
阅读原文 在惯性聚合中打开

推荐订阅源

博客园 - 叶小钗
Microsoft Azure Blog
Microsoft Azure Blog
Stack Overflow Blog
Stack Overflow Blog
Jina AI
Jina AI
Vercel News
Vercel News
H
Help Net Security
Martin Fowler
Martin Fowler
美团技术团队
云风的 BLOG
云风的 BLOG
Y
Y Combinator Blog
阮一峰的网络日志
阮一峰的网络日志
MyScale Blog
MyScale Blog
钛媒体:引领未来商业与生活新知
钛媒体:引领未来商业与生活新知
博客园 - 三生石上(FineUI控件)
博客园 - 司徒正美
人人都是产品经理
人人都是产品经理
Engineering at Meta
Engineering at Meta
G
Google Developers Blog
Blog — PlanetScale
Blog — PlanetScale
MongoDB | Blog
MongoDB | Blog
宝玉的分享
宝玉的分享
小众软件
小众软件
T
Tailwind CSS Blog
WordPress大学
WordPress大学

math updates on arXiv.org

Coupling-Robust Accuracy in Multiphysics Physics Informed Neural Networks via Kronecker-Preconditioned Optimization Non-normal spectral signatures of instability in neural network training dynamics Optimization of randomized neural networks for transfer operator approximation Selective Ambulance Dispatch Under Contextual Travel-Time Uncertainty LLAMA LIMA: A Living Meta-Analysis on the Effects of Generative AI on Learning Mathematics Neural Flow Operators can Approximate any Operator: Abstract Frameworks and Universal Approximations LLMs as Noisy Channels: A Shannon Perspective on Model Capacity and Scaling Laws On the Stability of Spherical Hellinger-Kantorovich Flows and Their Implications for Differential Privacy Training-Free Looped Transformers Move on Muon : A Hamiltonian probability gradient flow perspective of Muon optimizer Entrywise Error Bounds for Spectral Ranking with Semi-Random Adversaries Asymmetric Scaling Laws from Sparse Features Is Dimensionality a Barrier for Retrieval Models? RA-DCA: A Randomized Active-Set DCA for Directional Stationarity in Max-Structured DC Programs Commutator-Induced Uncertainty in VAEs Weisfeiler-Leman Is Incomplete on Simple Spectrum Graphs, so Canonicalize Them Sparse In-Network Learning via Shortest-Path Backpropagation and Finite-Rate Gating Instance-Optimal Estimation with Multiple LLM Judges on a Budget Entropy Equivalence Testing Expand More, Shrink Less: Shaping Effective-Rank Dynamics for Dense Scaling in Recommendation Any-Dimensional Invariant Universality Operationalizing Individual Fairness via Gradient Descent and Bradley-Terry Models Anytime Training with Schedule-Free Spectral Optimization Diffusion-based Denoising Beats Vanilla Score Matching in Parameter Estimation: A Theoretical Explanation Resilience Characterization of AI-Native Wireless Receivers via Persistent Homology The General Theory of Localization Methods Group-Algebraic Tensors: Provably-optimal Equivariant Learning and Physical Symmetry Discovery General Lower Bounds for Differentially Private Federated Learning with Arbitrary Public-Transcript Interactions PilotWiMAE: Pilot-Native Representation Learning for Wireless Channels Proximal basin hopping: global optimization with guarantees
Defect Spaces and Gram Operators for Tensor-Valued Incide...
Kengo Miyamoto · 2026-05-27 · via math updates on arXiv.org

We study vector-valued incidence maps obtained from ordinary graph incidence maps by linear observation of the free vertex space. Let $\mathbb{F}$ be a field, $D = (X, E, s, t)$ a finite directed multigraph, $U$ an $\mathbb{F}$-vector space, and $φ: X \to U$ a vertex labeling with $\mathbb{F}$-linear extension $\hatφ : \mathbb{F}^X \to U$. The vector-valued incidence map $\partial_φ: \mathbb{F}^E \to U$, $\partial_φ(\mathbf{1}_e) = φ(t(e)) - φ(s(e))$, factors as $\partial_φ= \hatφ \circ B_D$, where $B_D$ is the classical incidence map of $D$. We prove the formula $\dim_{\mathbb{F}} \mathrm{Ker}(\partial_φ) = |E| - |X| + c(D) + δ_φ,$ where $c(D)$ is the number of weakly connected components of $D$ and $δ_φ:= \dim_{\mathbb{F}}(\mathrm{Im}(B_D) \cap \mathrm{Ker}(\hatφ))$ is the defect invariant. We apply this framework to directed tensor-labeled hypergraphs $\mathcal{H} = (Q_0, Q_1, β)$, in which each hyperedge carries a pair of boundary tensors $(A_e, B_e)$ in the tensor algebra $T(\mathbb{F}^{Q_0})$, and prove that $δ(\mathcal{H}) = 0$ over any field for each of the six standard constructions, including symmetric encodings that degenerate in positive characteristic. Over $\mathbb{F} = \mathbb{R}$, the edge Gram operator $\mathcal{L} = \partial_β^* \partial_β$ has rank $|V_{\mathrm{macro}}| - c_{\mathrm{macro}} - δ(\mathcal{H})$, and its degree-truncated operators form a Loewner-monotone filtration whose rank increments equal the decrements of the defect filtration. We further realize the cycle space of every oriented hypergraph (in the sense of Reff--Rusnak) as $\mathrm{Ker}(\partial_β)$ within this framework, and exhibit a four-edge inclusion--exclusion example with $δ(\mathcal{H}) = 1$.