Benign Overfitting Does Not Occur in Diffusion Models

• URL: http://arxiv.org/abs/2607.02671v1
• Date: Thu, 02 Jul 2026 18:05:23 GMT
• Title: Benign Overfitting Does Not Occur in Diffusion Models
• Abstract summary: Diffusion models share much of the machinery of standard deep learning.<n>We show that, unless the sample size grows exponentially with the data dimension, overfitting and good generalization cannot occur simultaneously.<n>We identify implicit regularization stemming from time-smoothness of the score and early stopping during training as mechanisms that prevent such overfitting.
• Score: 33.8340803958907
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Benign overfitting and double descent have come to shape our understanding of generalization in deep learning, establishing that overfitting is not only compatible with good generalization but can actively benefit it. Diffusion models share much of the machinery of standard deep learning, so it is natural to assume that they also exhibit these properties. In this work, we show that this assumption is largely incorrect. We first establish fundamental impossibility results showing that, unless the sample size grows exponentially with the data dimension, overfitting and good generalization cannot occur simultaneously. Consequently, the population loss follows a classical U-shaped curve in model complexity rather than exhibiting double descent. Analyzing a simplified setting, we identify a key difference between regression and score matching: regression benefits from an alignment between the target and the empirical covariance; score matching admits no such alignment, leaving overfitting irreparably harmful. We further identify implicit regularization stemming from time-smoothness of the score and early stopping during training as mechanisms that prevent such overfitting and verify our findings with high-dimensional image generation experiments. Our results reveal that generalization in diffusion models is governed by mechanisms distinct from those of traditional regression, motivating the development of new theory.

Related papers

• (Mis)Understanding Benign Overfitting in Equity Return Prediction [14.203404489519242]
  We investigate whether this benign overfitting'' phenomenon extends to equity return prediction.<n>We find two key phenomena: first, a double descent pattern in the ridgeless model's prediction risk, and second, that while the optimal ridge model consistently outperforms its ridgeless counterpart, this performance gap becomes negligible at large parameter-to-observation ratios.
  arXiv  Detail & Related papers  (2026-08-24T18:54:34Z)
• Evaluating Double Descent in Machine Learning: Insights from Tree-Based Models Applied to a Genomic Prediction Task [0.0]
  Recent work has introduced the notion of a second descent in test error beyond the threshold-giving rise to the so-called double descent phenomenon.<n>We show that double descent consistently emerges only when complexity is scaled jointly across two axes.<n>Our findings underscore the importance of treating model complexity as a multidimensional construct when analysing generalisation behaviour.
  arXiv  Detail & Related papers  (2025-09-22T16:41:31Z)
• A Classical View on Benign Overfitting: The Role of Sample Size [14.36840959836957]
  We focus on almost benign overfitting, where models simultaneously achieve both arbitrarily small training and test errors.<n>This behavior is characteristic of neural networks, which often achieve low (but non-zero) training error while still generalizing well.
  arXiv  Detail & Related papers  (2025-05-16T18:37:51Z)
• On the Benefits of Over-parameterization for Out-of-Distribution Generalization [28.961538657831788]
  We investigate the performance of a machine learning model in terms of Out-of-Distribution (OOD) loss under benign overfitting conditions. We show that further increasing the model's parameterization can significantly reduce the OOD loss. These insights explain the empirical phenomenon of enhanced OOD generalization through model ensembles.
  arXiv  Detail & Related papers  (2024-03-26T11:01:53Z)
• A U-turn on Double Descent: Rethinking Parameter Counting in Statistical Learning [68.76846801719095]
  We show that double descent appears exactly when and where it occurs, and that its location is not inherently tied to the threshold p=n. This provides a resolution to tensions between double descent and statistical intuition.
  arXiv  Detail & Related papers  (2023-10-29T12:05:39Z)
• Nonparametric Identifiability of Causal Representations from Unknown Interventions [63.1354734978244]
  We study causal representation learning, the task of inferring latent causal variables and their causal relations from mixtures of the variables. Our goal is to identify both the ground truth latents and their causal graph up to a set of ambiguities which we show to be irresolvable from interventional data.
  arXiv  Detail & Related papers  (2023-06-01T10:51:58Z)
• Bias in Pruned Vision Models: In-Depth Analysis and Countermeasures [93.17009514112702]
  Pruning, setting a significant subset of the parameters of a neural network to zero, is one of the most popular methods of model compression. Despite existing evidence for this phenomenon, the relationship between neural network pruning and induced bias is not well-understood.
  arXiv  Detail & Related papers  (2023-04-25T07:42:06Z)
• Monotonicity and Double Descent in Uncertainty Estimation with Gaussian Processes [52.92110730286403]
  It is commonly believed that the marginal likelihood should be reminiscent of cross-validation metrics and that both should deteriorate with larger input dimensions. We prove that by tuning hyper parameters, the performance, as measured by the marginal likelihood, improves monotonically with the input dimension. We also prove that cross-validation metrics exhibit qualitatively different behavior that is characteristic of double descent.
  arXiv  Detail & Related papers  (2022-10-14T08:09:33Z)
• Ensembling over Classifiers: a Bias-Variance Perspective [13.006468721874372]
  We build upon the extension to the bias-variance decomposition by Pfau (2013) in order to gain crucial insights into the behavior of ensembles of classifiers. We show that conditional estimates necessarily incur an irreducible error. Empirically, standard ensembling reducesthe bias, leading us to hypothesize that ensembles of classifiers may perform well in part because of this unexpected reduction.
  arXiv  Detail & Related papers  (2022-06-21T17:46:35Z)
• Interpolation can hurt robust generalization even when there is no noise [76.3492338989419]
  We show that avoiding generalization through ridge regularization can significantly improve generalization even in the absence of noise. We prove this phenomenon for the robust risk of both linear regression and classification and hence provide the first theoretical result on robust overfitting.
  arXiv  Detail & Related papers  (2021-08-05T23:04:15Z)
• Optimization Variance: Exploring Generalization Properties of DNNs [83.78477167211315]
  The test error of a deep neural network (DNN) often demonstrates double descent. We propose a novel metric, optimization variance (OV), to measure the diversity of model updates.
  arXiv  Detail & Related papers  (2021-06-03T09:34:17Z)
• Understanding Double Descent Requires a Fine-Grained Bias-Variance Decomposition [34.235007566913396]
  We describe an interpretable, symmetric decomposition of the variance into terms associated with the labels. We find that the bias decreases monotonically with the network width, but the variance terms exhibit non-monotonic behavior. We also analyze the strikingly rich phenomenology that arises.
  arXiv  Detail & Related papers  (2020-11-04T21:04:02Z)
• Good Classifiers are Abundant in the Interpolating Regime [64.72044662855612]
  We develop a methodology to compute precisely the full distribution of test errors among interpolating classifiers. We find that test errors tend to concentrate around a small typical value $varepsilon*$, which deviates substantially from the test error of worst-case interpolating model. Our results show that the usual style of analysis in statistical learning theory may not be fine-grained enough to capture the good generalization performance observed in practice.
  arXiv  Detail & Related papers  (2020-06-22T21:12:31Z)

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.