An analytic theory of creativity in convolutional diffusion models

• URL: http://arxiv.org/abs/2412.20292v2
• Date: Thu, 05 Jun 2025 05:09:27 GMT
• Title: An analytic theory of creativity in convolutional diffusion models
• Authors: Mason Kamb, Surya Ganguli,
• Abstract summary: We obtain an analytic, interpretable and predictive theory of creativity in convolutional diffusion models.<n>Our model reveals a locally consistent patch mosaic mechanism of creativity.<n>Our theory also partially predicts the outputs of pre-trained self-attention enabled UNets.
• Score: 25.013838586673906
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: We obtain an analytic, interpretable and predictive theory of creativity in convolutional diffusion models. Indeed, score-matching diffusion models can generate highly original images that lie far from their training data. However, optimal score-matching theory suggests that these models should only be able to produce memorized training examples. To reconcile this theory-experiment gap, we identify two simple inductive biases, locality and equivariance, that: (1) induce a form of combinatorial creativity by preventing optimal score-matching; (2) result in fully analytic, completely mechanistically interpretable, local score (LS) and equivariant local score (ELS) machines that, (3) after calibrating a single time-dependent hyperparameter can quantitatively predict the outputs of trained convolution only diffusion models (like ResNets and UNets) with high accuracy (median $r^2$ of $0.95, 0.94, 0.94, 0.96$ for our top model on CIFAR10, FashionMNIST, MNIST, and CelebA). Our model reveals a locally consistent patch mosaic mechanism of creativity, in which diffusion models create exponentially many novel images by mixing and matching different local training set patches at different scales and image locations. Our theory also partially predicts the outputs of pre-trained self-attention enabled UNets (median $r^2 \sim 0.77$ on CIFAR10), revealing an intriguing role for attention in carving out semantic coherence from local patch mosaics.

Related papers

• Consistent World Models via Foresight Diffusion [56.45012929930605]
  We argue that a key bottleneck in learning consistent diffusion-based world models lies in the suboptimal predictive ability.<n>We propose Foresight Diffusion (ForeDiff), a diffusion-based world modeling framework that enhances consistency by decoupling condition understanding from target denoising.
  arXiv  Detail & Related papers  (2025-05-22T10:01:59Z)
• Generalization through variance: how noise shapes inductive biases in diffusion models [0.0]
  We develop a mathematical theory that partly explains 'generalization through variance' phenomenon. We find that the distributions diffusion models effectively learn to sample from resemble their training distributions. We also characterize how this inductive bias interacts with feature-related inductive biases.
  arXiv  Detail & Related papers  (2025-04-16T23:41:10Z)
• The Unreasonable Effectiveness of Gaussian Score Approximation for Diffusion Models and its Applications [1.8416014644193066]
  We compare learned neural scores to the scores of two kinds of analytically tractable distributions.<n>We claim that the learned neural score is dominated by its linear (Gaussian) approximation for moderate to high noise scales.<n>We show that this allows the skipping of the first 15-30% of sampling steps while maintaining high sample quality.
  arXiv  Detail & Related papers  (2024-12-12T21:31:27Z)
• How Diffusion Models Learn to Factorize and Compose [14.161975556325796]
  Diffusion models are capable of generating photo-realistic images that combine elements which likely do not appear together in the training set. We investigate whether and when diffusion models learn semantically meaningful and factorized representations of composable features.
  arXiv  Detail & Related papers  (2024-08-23T17:59:03Z)
• Provable Statistical Rates for Consistency Diffusion Models [87.28777947976573]
  Despite the state-of-the-art performance, diffusion models are known for their slow sample generation due to the extensive number of steps involved. This paper contributes towards the first statistical theory for consistency models, formulating their training as a distribution discrepancy minimization problem.
  arXiv  Detail & Related papers  (2024-06-23T20:34:18Z)
• BEND: Bagging Deep Learning Training Based on Efficient Neural Network Diffusion [56.9358325168226]
  We propose a Bagging deep learning training algorithm based on Efficient Neural network Diffusion (BEND) Our approach is simple but effective, first using multiple trained model weights and biases as inputs to train autoencoder and latent diffusion model. Our proposed BEND algorithm can consistently outperform the mean and median accuracies of both the original trained model and the diffused model.
  arXiv  Detail & Related papers  (2024-03-23T08:40:38Z)
• Theoretical Insights for Diffusion Guidance: A Case Study for Gaussian Mixture Models [59.331993845831946]
  Diffusion models benefit from instillation of task-specific information into the score function to steer the sample generation towards desired properties. This paper provides the first theoretical study towards understanding the influence of guidance on diffusion models in the context of Gaussian mixture models.
  arXiv  Detail & Related papers  (2024-03-03T23:15:48Z)
• Analyzing Neural Network-Based Generative Diffusion Models through Convex Optimization [45.72323731094864]
  We present a theoretical framework to analyze two-layer neural network-based diffusion models. We prove that training shallow neural networks for score prediction can be done by solving a single convex program. Our results provide a precise characterization of what neural network-based diffusion models learn in non-asymptotic settings.
  arXiv  Detail & Related papers  (2024-02-03T00:20:25Z)
• The Hidden Linear Structure in Score-Based Models and its Application [2.1756081703276]
  We show that for well-trained diffusion models, the learned score at a high noise scale is well approximated by the linear score of Gaussian. Our finding of the linear structure in the score-based model has implications for better model design and data pre-processing.
  arXiv  Detail & Related papers  (2023-11-17T22:25:07Z)
• Discrete Diffusion Modeling by Estimating the Ratios of the Data Distribution [67.9215891673174]
  We propose score entropy as a novel loss that naturally extends score matching to discrete spaces. We test our Score Entropy Discrete Diffusion models on standard language modeling tasks.
  arXiv  Detail & Related papers  (2023-10-25T17:59:12Z)
• Towards Faster Non-Asymptotic Convergence for Diffusion-Based Generative Models [49.81937966106691]
  We develop a suite of non-asymptotic theory towards understanding the data generation process of diffusion models. In contrast to prior works, our theory is developed based on an elementary yet versatile non-asymptotic approach.
  arXiv  Detail & Related papers  (2023-06-15T16:30:08Z)
• Diff-Instruct: A Universal Approach for Transferring Knowledge From Pre-trained Diffusion Models [77.83923746319498]
  We propose a framework called Diff-Instruct to instruct the training of arbitrary generative models. We show that Diff-Instruct results in state-of-the-art single-step diffusion-based models. Experiments on refining GAN models show that the Diff-Instruct can consistently improve the pre-trained generators of GAN models.
  arXiv  Detail & Related papers  (2023-05-29T04:22:57Z)
• Your Diffusion Model is Secretly a Zero-Shot Classifier [90.40799216880342]
  We show that density estimates from large-scale text-to-image diffusion models can be leveraged to perform zero-shot classification. Our generative approach to classification attains strong results on a variety of benchmarks. Our results are a step toward using generative over discriminative models for downstream tasks.
  arXiv  Detail & Related papers  (2023-03-28T17:59:56Z)
• Understanding Diffusion Models: A Unified Perspective [0.0]
  Diffusion models have shown incredible capabilities as generative models. We review, demystify, and unify the understanding of diffusion models across both variational and score-based perspectives.
  arXiv  Detail & Related papers  (2022-08-25T09:55:25Z)
• How Much is Enough? A Study on Diffusion Times in Score-based Generative Models [76.76860707897413]
  Current best practice advocates for a large T to ensure that the forward dynamics brings the diffusion sufficiently close to a known and simple noise distribution. We show how an auxiliary model can be used to bridge the gap between the ideal and the simulated forward dynamics, followed by a standard reverse diffusion process.
  arXiv  Detail & Related papers  (2022-06-10T15:09:46Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.