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

Google DeepMind News
Google DeepMind News
Attack and Defense Labs
Attack and Defense Labs
GbyAI
GbyAI
月光博客
月光博客
Recent Announcements
Recent Announcements
云风的 BLOG
云风的 BLOG
F
Full Disclosure
宝玉的分享
宝玉的分享
freeCodeCamp Programming Tutorials: Python, JavaScript, Git & More
爱范儿
爱范儿
博客园 - Franky
V
V2EX
Recorded Future
Recorded Future
WordPress大学
WordPress大学
小众软件
小众软件
Webroot Blog
Webroot Blog
雷峰网
雷峰网
Vercel News
Vercel News
N
News and Events Feed by Topic
PCI Perspectives
PCI Perspectives
Cyberwarzone
Cyberwarzone
Threat Intelligence Blog | Flashpoint
Threat Intelligence Blog | Flashpoint
W
WeLiveSecurity
S
Schneier on Security
博客园 - 三生石上(FineUI控件)
K
Kaspersky official blog
F
Fortinet All Blogs
Security Archives - TechRepublic
Security Archives - TechRepublic
Spread Privacy
Spread Privacy
博客园 - 叶小钗
罗磊的独立博客
D
Docker
Forbes - Security
Forbes - Security
让小产品的独立变现更简单 - ezindie.com
让小产品的独立变现更简单 - ezindie.com
The Register - Security
The Register - Security
A
About on SuperTechFans
B
Blog RSS Feed
I
InfoQ
T
Tailwind CSS Blog
G
Google Developers Blog
H
Help Net Security
V
Vulnerabilities – Threatpost
AWS News Blog
AWS News Blog
N
News and Events Feed by Topic
M
MIT News - Artificial intelligence
K
KPMG report finds enterprise disconnect between AI and its ROI | CIO
P
Proofpoint News Feed
cs.AI updates on arXiv.org
cs.AI updates on arXiv.org
MyScale Blog
MyScale Blog
Latest news
Latest 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 Learning Decision-Sufficient Representations for Linear Optimization Parameterized Complexity of Stationarity Testing for Piecewise-Affine Functions and Shallow CNN Losses Prabhakar function and unified fractional kinetic equation in bicomplex space Computing Gamma(p/q) with Beta function values Flows on Graded Manifolds Optimal embedding dimension in the Nash--Tognoli theorem An optimal first-order method for smooth and strongly convex composite optimization and its stationary limit Sharp Bohr-Type inequalities for certain classes of close-to-convex functions Invariants of real affine varieties based on their complexifications Topological symmetric and braid homologies A Formal Graph-Theoretic Framework for Pitch Class Set Analysis Finite groups with high commuting probability for Sylow subgroups Performance Bounds for Rollout Policies in Stochastic Shortest Path Problems Real 2-blocks in quasi-simple groups Maximal subalgebras of the Lie algebra $W_n(\mathbb{K})$ Cohomogeneity-One Ruled Hypersurfaces in $\mathbb{CP}^2$ and $\mathbb{C}H^2$ 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 Generalized Stochastic Approximation of the Log-Likelihood Ratio for Robust Sequential Change-Point Detection 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 Concise and elegant proofs of three formulas for complete Bell polynomials On Reed-Muller subcodes, Grassmannian partitions and sum-free functions 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 A Comprehensive Study of Clique Graphs and Clique Regular Graphs Every signed planar graph is $5$-choosable: A short proof and refinements 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 Democratizing Large-Scale Re-Optimization with LLM-Guided Model Patches On Stability and Decomposition of Sample Quantiles under Heavy-Tailed Distributions Symmetry-Compatible Principle for Optimizer Design: Embeddings, LM Heads, SwiGLU MLPs, and MoE Routers Stochastic Non-Smooth Convex Optimization with Unbounded Gradients Dimension-Free Convergence of Discrete Diffusion Models: Adjoint Equations Induce the Right Space The Geometry of Cooperative Game Solutions: Stratified Egalitarian Shapley Values An Axiomatic Theory of Tie-Breaking: Impossibility, Characterization, and Decomposition Amortized Energy-Based Bayesian Inference PyCSP3-Scheduling: A Scheduling Extension for PyCSP3 Strategic PAC Learnability via Geometric Definability Proximal-Based Generative Modeling for Bayesian Inverse Problems Every Minimal Counterexample to the Erdős-Gyárfás Conjecture is Predominantly Cubic SPHERICAL KV: Angle-Domain Attention and Rate-Distortion Retention for Efficient Long-Context Inference NOVA: Fundamental Limits of Knowledge Discovery Through AI Model-based Bootstrap of Controlled Markov Chains TopoGeoScore: A Self-Supervised Source-Only Geometric Framework for OOD Checkpoint Selection Minimal Filling Architectures of Polynomial Neural Networks: Counterexamples, Frontier Search, and Defects Omni-scale Learning-based Sequential Decision Framework for Order Fulfillment of Tote-handling Robotic Systems Grokking or Glitching? How Low-Precision Drives Slingshot Loss Spikes Towards an Inferentialist Account of Information Through Proof-theoretic Semantics Random test functions, $H^{-1}$ norm equivalence, and stochastic variational physics-informed neural networks QUIVER: Cost-Aware Adaptive Preference Querying in Surrogate-Assisted Evolutionary Multi-Objective Optimization Robust and Fast Training via Per-Sample Clipping Beyond Continuity: Simulation-free Reconstruction of Discrete Branching Dynamics from Single-cell Snapshots Wasserstein Distributionally Robust Regret Optimization for Reinforcement Learning from Human Feedback Deep Learning of Solver-Aware Turbulence Closures from Nudged LES Dynamics Information bottleneck for learning the phase space of dynamics from high-dimensional experimental data QED: An Open-Source Multi-Agent System for Generating Mathematical Proofs on Open Problems Information-Theoretic Measures in AI: A Practical Decision Guide Inference of Online Newton Methods with Nesterov's Accelerated Sketching A Unified Fractional Regularization Framework for Sparse Recovery Mathematical Foundations for Peer-to-Peer Lattice Computation Geometric Layer-wise Approximation Rates for Deep Networks RateQuant: Optimal Mixed-Precision KV Cache Quantization via Rate-Distortion Theory ML-based approach to classification and generation of structured light propagation in turbulent media Zeroth-Order Optimization at the Edge of Stability Adaptive Learning via Off-Model Training and Importance Sampling for Fully Non-Markovian Optimal Stochastic Control. Complete version Beyond Fixed False Discovery Rates: Post-Hoc Conformal Selection with E-Variables Order-Optimal Sequential 1-Bit Mean Estimation in General Tail Regimes Training-Free Rate-Distortion-Perception Traversal With Diffusion Conformal Policy Control Linear Regression with Unknown Truncation Beyond Gaussian Features ArcMark: Distortion-Free Multi-Byte LLM Watermark via Optimal Transport Feature Learning Dynamics in Infinite-Depth Neural Networks ATHENA: Agentic Team for Hierarchical Evolutionary Numerical Algorithms Normalizing Flows on Quotient Manifolds via Boundary Quotients What Can Be Recovered Under Sparse Adversarial Corruption? Assumption-Free Theory for Linear Measurements TelecomTS: A Multi-Modal Observability Dataset for Time Series and Language Analysis Program Evaluation with Remotely Sensed Outcomes Efficient Gradient Estimation for Parameterized Quantum Systems with Lie Algebraic Symmetries
Global Convergence of Gradient Descent for Score Matching in Gaussian Mixtures via Reverse Fisher Divergence
[Submitted on 18 Jun 2026] · 2026-06-19 · via math updates on arXiv.org

Computer Science > Machine Learning

arXiv:2606.19876 (cs)

[Submitted on 18 Jun 2026]

View PDF

Abstract:The score matching problem is a central training objective in modern generative modeling, diffusion models, fitting unnormalized statistical models, and inverse problems. A standard approach is to minimize the forward Fisher divergence, where the expectation is taken with respect to the teacher distribution. However, recent results show that even in simple Gaussian mixture model settings, this objective can lead to undesirable and initialization-dependent convergence behavior. In this paper, we study an alternative objective: the reverse Fisher divergence, where the expectation is taken with respect to the student distribution. We analyze gradient descent (GD) for fitting Gaussian mixture models and show that this change in the objective leads to significantly better optimization properties. First, when the teacher distribution is a single Gaussian and the student is a Gaussian mixture model with fixed weights and identity covariances, we prove the global convergence of GD from arbitrary initializations. Second, we extend the analysis to the case where the teacher is also a Gaussian mixture model and prove global convergence guarantees under a global random initialization scheme and a $\widetilde{\Omega}(1)$-separation assumption on the target means. In particular, with high probability, each student component converges near its closest teacher component, and we provide conditions under which the student distribution converges in total variation distance. Our proofs rely on a new Lyapunov-based analysis of the gradient descent dynamics, showing that the reverse Fisher divergence has a much more favorable optimization landscape than the forward Fisher divergence.

Submission history

From: Alexander Tyurin [view email]
[v1] Thu, 18 Jun 2026 07:34:33 UTC (105 KB)

Current browse context:

cs.LG

Bookmark

BibSonomy Reddit

Bibliographic Tools

Bibliographic and Citation Tools

Bibliographic Explorer Toggle

Code, Data, Media

Code, Data and Media Associated with this Article

Demos

Demos

Related Papers

Recommenders and Search Tools

IArxiv recommender toggle

About arXivLabs

arXivLabs: experimental projects with community collaborators

arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.

Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.

Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.