{"total":104,"rows":[{"id":"2505.17638","title":"Why Diffusion Models Don't Memorize: The Role of Implicit Dynamical Regularization in Training","created":"2025-05-23","cited_by_count":81,"area":"diffusion"},{"id":"2506.03719","title":"On the Closed-Form of Flow Matching: Generalization Does Not Arise from Target Stochasticity","created":"2025-06-04","cited_by_count":50,"area":"diffusion"},{"id":"2505.21777","title":"Memorization to Generalization: Emergence of Diffusion Models from Associative Memory","created":"2025-05-27","cited_by_count":36,"area":"diffusion"},{"id":"2504.15266","title":"Roll the dice & look before you leap: Going beyond the creative limits of next-token prediction","created":"2025-04-21","cited_by_count":29,"area":null},{"id":"2604.21691","title":"There Will Be a Scientific Theory of Deep Learning","created":"2026-04-23","cited_by_count":29,"area":null},{"id":"2503.03206","title":"An Analytical Theory of Spectral Bias in the Learning Dynamics of Diffusion Models","created":"2025-03-05","cited_by_count":26,"area":"diffusion"},{"id":"2502.12089","title":"How Compositional Generalization and Creativity Improve as Diffusion Models are Trained","created":"2025-02-17","cited_by_count":25,"area":"diffusion"},{"id":"2505.16959","title":"Bigger Isn't Always Memorizing: Early Stopping Overparameterized Diffusion Models","created":"2025-05-22","cited_by_count":25,"area":"diffusion"},{"id":"2509.09672","title":"Locality in Image Diffusion Models Emerges from Data Statistics","created":"2025-09-11","cited_by_count":25,"area":"diffusion"},{"id":"2504.12532","title":"Generalization through variance: how noise shapes inductive biases in diffusion models","created":"2025-04-16","cited_by_count":23,"area":"diffusion"},{"id":"2508.17689","title":"On the Edge of Memorization in Diffusion Models","created":"2025-08-25","cited_by_count":22,"area":"diffusion"},{"id":"2502.00336","title":"Denoising Score Matching with Random Features: Insights on Diffusion Models from Precise Learning Curves","created":"2025-02-01","cited_by_count":20,"area":"diffusion"},{"id":"2510.02305","title":"Diffusion Models and the Manifold Hypothesis: Log-Domain Smoothing is Geometry Adaptive","created":"2025-10-02","cited_by_count":20,"area":"diffusion"},{"id":"2502.05743","title":"Understanding Representation Dynamics of Diffusion Models via Low-Dimensional Modeling","created":"2025-02-09","cited_by_count":18,"area":"diffusion"},{"id":"2506.10038","title":"Ambient Diffusion Omni: Training Good Models with Bad Data","created":"2025-06-10","cited_by_count":18,"area":"diffusion"},{"id":"2509.16499","title":"A Closer Look at Model Collapse: From a Generalization-to-Memorization Perspective","created":"2025-09-20","cited_by_count":18,"area":"diffusion"},{"id":"2510.01378","title":"Selective Underfitting in Diffusion Models","created":"2025-10-01","cited_by_count":18,"area":"diffusion"},{"id":"2502.21278","title":"Does Generation Require Memorization? Creative Diffusion Models using Ambient Diffusion","created":"2025-02-28","cited_by_count":17,"area":"diffusion"},{"id":"2502.19499","title":"On the Interpolation Effect of Score Smoothing in Diffusion Models","created":"2025-02-26","cited_by_count":15,"area":"diffusion"},{"id":"2501.03937","title":"A solvable model of learning generative diffusion: theory and insights","created":"2025-01-07","cited_by_count":14,"area":"diffusion"},{"id":"2502.04549","title":"Mechanisms of Projective Composition of Diffusion Models","created":"2025-02-06","cited_by_count":14,"area":"diffusion"},{"id":"2511.03202","title":"Provable Separations between Memorization and Generalization in Diffusion Models","created":"2025-11-05","cited_by_count":14,"area":"diffusion"},{"id":"2511.19428","title":"Flow Map Distillation Without Data","created":"2025-11-24","cited_by_count":14,"area":null},{"id":"2502.02300","title":"Density Ratio Estimation with Conditional Probability Paths","created":"2025-02-04","cited_by_count":13,"area":null},{"id":"2603.05630","title":"Making Reconstruction FID Predictive of Diffusion Generation FID","created":"2026-03-05","cited_by_count":13,"area":"diffusion"},{"id":"2512.20963","title":"Generalization of Diffusion Models Arises with a Balanced Representation Space","created":"2025-12-24","cited_by_count":11,"area":"diffusion"},{"id":"2505.02508","title":"Resolving Memorization in Empirical Diffusion Model for Manifold Data in High-Dimensional Spaces","created":"2025-05-05","cited_by_count":8,"area":"diffusion"},{"id":"2508.06614","title":"Local Diffusion Models and Phases of Data Distributions","created":"2025-08-08","cited_by_count":8,"area":"diffusion"},{"id":"2512.02826","title":"From Navigation to Refinement: Revealing the Two-Stage Nature of Flow-based Diffusion Models through Oracle Velocity","created":"2025-12-02","cited_by_count":8,"area":"diffusion"},{"id":"2510.24830","title":"The Generation Phases of Flow Matching: a Denoising Perspective","created":"2025-10-28","cited_by_count":7,"area":"diffusion"},{"id":"2601.21419","title":"Revisiting Diffusion Model Predictions Through Dimensionality","created":"2026-01-29","cited_by_count":7,"area":"diffusion"},{"id":"2602.02908","title":"A Random Matrix Theory Perspective on the Consistency of Diffusion Models","created":"2026-02-02","cited_by_count":6,"area":"diffusion"},{"id":"2603.20092","title":"How Out-of-Equilibrium Phase Transitions can Seed Pattern Formation in Trained Diffusion Models","created":"2026-03-20","cited_by_count":6,"area":"diffusion"},{"id":"2505.04417","title":"Localized Diffusion Models","created":"2025-05-07","cited_by_count":5,"area":"diffusion"},{"id":"2510.25976","title":"Brain-IT: Image Reconstruction from fMRI via Brain-Interaction Transformer","created":"2025-10-29","cited_by_count":5,"area":"diffusion"},{"id":"2602.06021","title":"Diffusion Model's Generalization Can Be Characterized by Inductive Biases toward a Data-Dependent Ridge Manifold","created":"2026-02-05","cited_by_count":5,"area":"diffusion"},{"id":"2606.14390","title":"Local Coverage Governs Memorization in Diffusion Models","created":"2026-06-12","cited_by_count":5,"area":"diffusion"},{"id":"2501.18797","title":"Compositional Generalization via Forced Rendering of Disentangled Latents","created":"2025-01-30","cited_by_count":4,"area":null},{"id":"2506.10177","title":"Geometric Regularity in Deterministic Sampling Dynamics of Diffusion-based Generative Models","created":"2025-06-11","cited_by_count":4,"area":"diffusion"},{"id":"2506.17324","title":"Origins of Creativity in Attention-Based Diffusion Models","created":"2025-06-18","cited_by_count":4,"area":"diffusion"},{"id":"2506.19136","title":"Local Learning Rules for Out-of-Equilibrium Physical Generative Models","created":"2025-06-23","cited_by_count":4,"area":"diffusion"},{"id":"2510.14269","title":"Nonparametric Data Attribution for Diffusion Models","created":"2025-10-16","cited_by_count":4,"area":"diffusion"},{"id":"2510.26997","title":"Gradient Descent as Loss Landscape Navigation: a Normative Framework for Deriving Learning Rules","created":"2025-10-30","cited_by_count":4,"area":null},{"id":"2603.03469","title":"Biased Generalization in Diffusion Models","created":"2026-03-03","cited_by_count":4,"area":"diffusion"},{"id":"2504.13333","title":"Predicting Forced Responses of Probability Distributions via the Fluctuation-Dissipation Theorem and Generative Modeling","created":"2025-04-17","cited_by_count":3,"area":null},{"id":"2505.22841","title":"Kernel-Smoothed Scores for Denoising Diffusion: A Bias-Variance Study","created":"2025-05-28","cited_by_count":3,"area":"diffusion"},{"id":"2509.16447","title":"Local Mechanisms of Compositional Generalization in Conditional Diffusion","created":"2025-09-19","cited_by_count":3,"area":"diffusion"},{"id":"2509.25003","title":"Score-based Membership Inference on Diffusion Models","created":"2025-09-29","cited_by_count":3,"area":"diffusion"},{"id":"2510.03075","title":"What Drives Compositional Generalization in Visual Generative Models? The Importance of Continuous Training Objectives","created":"2025-10-03","cited_by_count":3,"area":null},{"id":"2602.17846","title":"Two Calm Ends and the Wild Middle: A Geometric Picture of Memorization in Diffusion Models","created":"2026-02-19","cited_by_count":3,"area":"diffusion"},{"id":"2602.20360","title":"Momentum Guidance: Plug-and-Play Guidance for Flow Models","created":"2026-02-23","cited_by_count":3,"area":null},{"id":"2607.08041","title":"An exact information theory of generalization phase transitions in Bayesian diffusion models","created":"2026-07-09","cited_by_count":3,"area":"diffusion"},{"id":"2608.23938","title":"Generalization, memorization, and overfitting for diffusion models trained in the lazy high-dimensional regime","created":"2026-08-25","cited_by_count":3,"area":"diffusion"},{"id":"2501.07376","title":"Bigger Isn't Always Better: Towards a General Prior for Medical Image Reconstruction","created":"2025-01-13","cited_by_count":2,"area":"diffusion"},{"id":"2505.20123","title":"Understanding Generalization in Diffusion Distillation via Probability Flow Distance","created":"2025-05-26","cited_by_count":2,"area":"diffusion"},{"id":"2509.21470","title":"Score-based Idempotent Distillation of Diffusion Models","created":"2025-09-25","cited_by_count":2,"area":"diffusion"},{"id":"2601.10334","title":"An analytic theory of convolutional neural network inverse problems solvers","created":"2026-01-15","cited_by_count":2,"area":null},{"id":"2601.15417","title":"Ambient Dataloops: Generative Models for Dataset Refinement","created":"2026-01-21","cited_by_count":2,"area":"diffusion"},{"id":"2603.13419","title":"Diffusion Models Memorize in Training -- and Generalize in Inference","created":"2026-03-12","cited_by_count":2,"area":"diffusion"},{"id":"2605.04830","title":"Concurrence of Symmetry Breaking and Nonlocality Phase Transitions in Diffusion Models","created":"2026-05-06","cited_by_count":2,"area":"diffusion"},{"id":"2605.06077","title":"Understanding diffusion models requires rethinking (again) generalization","created":"2026-05-07","cited_by_count":2,"area":"diffusion"},{"id":"2605.16396","title":"Beyond MMSE: Enhancing PnP Restoration with ProxiMAP","created":"2026-05-12","cited_by_count":2,"area":null},{"id":"2503.05558","title":"Diffusion Models for Cayley Graphs","created":"2025-03-07","cited_by_count":1,"area":"diffusion"},{"id":"2506.00814","title":"GuessBench: Sensemaking Multimodal Creativity in the Wild","created":"2025-06-01","cited_by_count":1,"area":"mllm"},{"id":"2506.01912","title":"Unconditional CNN denoisers contain sparse semantic representation of images","created":"2025-06-02","cited_by_count":1,"area":"diffusion"},{"id":"2509.22318","title":"NIFTY: a Non-Local Image Flow Matching for Texture Synthesis","created":"2025-09-26","cited_by_count":1,"area":"diffusion"},{"id":"2601.00781","title":"Categorical Reparameterization with Denoising Diffusion models","created":"2026-01-02","cited_by_count":1,"area":"diffusion"},{"id":"2605.00273","title":"When Do Diffusion Models learn to Generate Multiple Objects?","created":"2026-04-30","cited_by_count":1,"area":"diffusion"},{"id":"2605.10019","title":"The two clocks and the innovation window: When and how generative models learn rules","created":"2026-05-11","cited_by_count":1,"area":null},{"id":"2605.12597","title":"The critical slowing down in training diffusion models","created":"2026-05-12","cited_by_count":1,"area":"diffusion"},{"id":"2605.21402","title":"Memorisation, convergence and generalisation in generative models","created":"2026-05-20","cited_by_count":1,"area":null},{"id":"2605.30642","title":"Diffusion Models Preferentially Memorize Prototypical Examples or: Why Does My Diffusion Model Love Slop?","created":"2026-05-28","cited_by_count":1,"area":"diffusion"},{"id":"2606.04299","title":"Efficient and Training-Free Single-Image Diffusion Models","created":"2026-06-03","cited_by_count":1,"area":"diffusion"},{"id":"2606.08309","title":"Where the Score Lives: A Wavelet View of Diffusion","created":"2026-06-06","cited_by_count":1,"area":"diffusion"},{"id":"2606.09718","title":"Evaluating the Representation Space of Diffusion Models via Self-Supervised Principles","created":"2026-06-08","cited_by_count":1,"area":"diffusion"},{"id":"2609.00885","title":"Denoising Diffusion Generative Models Secretly Calculate Attentions","created":"2026-09-01","cited_by_count":1,"area":"diffusion"},{"id":"2503.04687","title":"Compositional World Knowledge leads to High Utility Synthetic data","created":"2025-03-06","cited_by_count":0,"area":null},{"id":"2505.18083","title":"What Do You Need for Compositional Generalization in Diffusion Planning?","created":"2025-05-23","cited_by_count":0,"area":"offlinerl"},{"id":"2507.04709","title":"Spooky Action at a Distance: Normalization Layers Enable Side-Channel Spatial Communication","created":"2025-07-07","cited_by_count":0,"area":null},{"id":"2510.22196","title":"Scaling Non-Parametric Sampling with Representation","created":"2025-10-25","cited_by_count":0,"area":null},{"id":"2512.11415","title":"Emergence of Nonequilibrium Latent Cycles in Unsupervised Generative Modeling","created":"2025-12-12","cited_by_count":0,"area":null},{"id":"2602.02699","title":"Sparsely Supervised Diffusion","created":"2026-02-02","cited_by_count":0,"area":"diffusion"},{"id":"2602.16498","title":"Fast and Scalable Analytical Diffusion","created":"2026-02-18","cited_by_count":0,"area":"diffusion"},{"id":"2602.19028","title":"The Metaphysics We Train: A Heideggerian Reading of Machine Learning","created":"2025-11-25","cited_by_count":0,"area":null},{"id":"2603.20987","title":"Interpreting the Synchronization Gap: The Hidden Mechanism Inside Diffusion Transformers","created":"2026-03-22","cited_by_count":0,"area":"diffusion"},{"id":"2604.15306","title":"Generalization in LLM Problem Solving: The Case of the Shortest Path","created":"2026-04-16","cited_by_count":0,"area":null},{"id":"2604.17673","title":"Grokking of Diffusion Models: Case Study on Modular Addition","created":"2026-04-20","cited_by_count":0,"area":"diffusion"},{"id":"2604.26841","title":"Language Diffusion Models are Associative Memories Capable of Retrieving Unseen Data","created":"2026-04-29","cited_by_count":0,"area":"diffusion"},{"id":"2605.14276","title":"Training-Free Generative Sampling via Moment-Matched Score Smoothing","created":"2026-05-14","cited_by_count":0,"area":"diffusion"},{"id":"2605.16348","title":"Flow-Direct: Feedback-Efficient and Reusable Guidance for Flow Models via Non-Parametric Guidance Field","created":"2026-05-08","cited_by_count":0,"area":null},{"id":"2605.16415","title":"Diffusion Models, Denoiser Architecture and Creativity","created":"2026-05-13","cited_by_count":0,"area":"diffusion"},{"id":"2605.19554","title":"Self-Creative Text-to-Object Generation using Semantic-Aware Spatial Weighting","created":"2026-05-19","cited_by_count":0,"area":null},{"id":"2605.20299","title":"Mechanisms of Misgeneralization in Physical Sequence Modeling","created":"2026-05-19","cited_by_count":0,"area":null},{"id":"2605.21692","title":"Representation Gap: Explaining the Unreasonable Effectiveness of Neural Networks from a Geometric Perspective","created":"2026-05-20","cited_by_count":0,"area":null},{"id":"2605.24192","title":"Filtered Posterior Mean Collections: A Unified Framework for Analytical Models of Diffusion Generalization","created":"2026-05-22","cited_by_count":0,"area":"diffusion"},{"id":"2606.01238","title":"Training-Free Imitation Learning with Closed-Form Diffusion Policies","created":"2026-05-31","cited_by_count":0,"area":"diffusion"},{"id":"2607.15693","title":"Toward a mechanistic understanding of inference in visual cortex and diffusion models","created":"2026-07-17","cited_by_count":0,"area":"diffusion"},{"id":"2608.14308","title":"Information Spreading in Diffusion Models from Effective Field Theory","created":"2026-08-14","cited_by_count":0,"area":"diffusion"},{"id":"2608.23696","title":"Renormalization Group Flow Matching for Scalable Local Generative Modeling","created":"2026-08-24","cited_by_count":0,"area":"diffusion"},{"id":"2609.32432","title":"Beyond the Manifold Hypothesis: Hybrid Spectral Parameterizations for Flow Matching","created":"2026-09-26","cited_by_count":0,"area":"diffusion"}],"next_cursor":"100"}