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

Last Week in AI
Last Week in AI
H
Help Net Security
博客园 - 叶小钗
V
Visual Studio Blog
阮一峰的网络日志
阮一峰的网络日志
博客园 - 三生石上(FineUI控件)
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
Microsoft Azure Blog
Microsoft Azure Blog
G
Google Developers Blog
腾讯CDC
MongoDB | Blog
MongoDB | Blog
宝玉的分享
宝玉的分享
P
Proofpoint News Feed
GbyAI
GbyAI
Microsoft Security Blog
Microsoft Security Blog
A
About on SuperTechFans
博客园 - 司徒正美
人人都是产品经理
人人都是产品经理
T
The Blog of Author Tim Ferriss
Martin Fowler
Martin Fowler
月光博客
月光博客
博客园 - 聂微东
酷 壳 – CoolShell
酷 壳 – CoolShell
Vercel News
Vercel News

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
A General Lower Bound for the Limited Augmented Zarankiew...
Liqun Qi, Chunfeng Cui, Yi Xu · 2026-04-05 · via math updates on arXiv.org

The limited augmented Zarankiewicz number $z_L(m,n)$ satisfies $\operatorname{BSR}(m,n)\ge z_L(m,n)\ge z(m,n)$, where $\operatorname{BSR}(m,n)$ is the maximum SOS rank of $m\times n$ biquadratic forms and $z(m,n)$ is the classical Zarankiewicz number. Our main result is a general lower bound for $z_L(m,n)$ based on the incidence graph of the complete graph $K_{4t}$. For every integer $t\ge 1$, let $m = \binom{4t}{2}$ and $n = 4t$. Then $$ \operatorname{BSR}(m,n) \;\ge\; z_L(m,n) \;\ge\; 2\binom{4t}{2} + 4t^2 - 2t. $$ Since $z = 2\binom{4t}{2} = Θ(t^2)$, the gap satisfies $z_L - z \ge 4t^2 - 2t = Θ(t^2) = Θ(m)$, i.e., it grows linearly in $m$. Moreover, $$ \frac{z_L - z}{z} \;\ge\; \frac{4t^2}{16t^2 - 4t} \;\longrightarrow\; \frac{1}{4} \quad \text{as } t\to\infty, $$ so the gap is asymptotically at least $25\%$ of $z$ -- a non-negligible constant fraction. For $t=1$ we obtain $z_L(6,4)\ge 14$, and we prove that this bound is tight, i.e., $z_L(6,4)=14$. For $t=2$ and $t=3$ we obtain $z_L(28,8)\ge 68$ and $z_L(66,12)\ge 162$, respectively, improving previously known bounds. We also determine the exact values of $z_L(m,n)$ for $5\times3$ and $5\times4$: $z_L(5,3)=9$ and $z_L(5,4)=12$. These results serve as base cases for a \emph{lifting method} that constructs admissible limited augmented graphs on $(m+1)\times(n+1)$ from optimal ones on $m\times n$. Applying this method, we obtain new lower bounds: $z_L(6,3)\ge 10$ and $z_L(6,5)\ge 17$. For $5\times 5$ we establish a new lower bound $z_L(5,5)\ge 15$, improving the previously known bound. By a direct construction, we have $z_L(6,6)\ge 20$.