Provably Reliable Classifier Guidance via Cross-Entropy Control

• URL: http://arxiv.org/abs/2601.21200v2
• Date: Thu, 05 Feb 2026 06:45:59 GMT
• Title: Provably Reliable Classifier Guidance via Cross-Entropy Control
• Authors: Sharan Sahu, Arisina Banerjee, Yuchen Wu,
• Abstract summary: We show that cross-entropy loss at each diffusion model step is sufficient to control the corresponding guidance error.<n>Our result yields an upper bound on the sampling error vectors-guided diffusion models and bears resemblance to a reverse log-Sobolev-type inequality.
• Score: 4.298880233819988
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Classifier-guided diffusion models generate conditional samples by augmenting the reverse-time score with the gradient of the log-probability predicted by a probabilistic classifier. In practice, this classifier is usually obtained by minimizing an empirical loss function. While existing statistical theory guarantees good generalization performance when the sample size is sufficiently large, it remains unclear whether such training yields an effective guidance mechanism. We study this question in the context of cross-entropy loss, which is widely used for classifier training. Under mild smoothness assumptions on the classifier, we show that controlling the cross-entropy at each diffusion model step is sufficient to control the corresponding guidance error. In particular, probabilistic classifiers achieving conditional KL divergence $\varepsilon^2$ induce guidance vectors with mean squared error $\widetilde O(d \varepsilon )$, up to constant and logarithmic factors. Our result yields an upper bound on the sampling error of classifier-guided diffusion models and bears resemblance to a reverse log-Sobolev--type inequality. To the best of our knowledge, this is the first result that quantitatively links classifier training to guidance alignment in diffusion models, providing both a theoretical explanation for the empirical success of classifier guidance, and principled guidelines for selecting classifiers that induce effective guidance.

Related papers

• Studying Classifier(-Free) Guidance From a Classifier-Centric Perspective [100.54185280153753]
  We find that both classifier guidance and classifier-free guidance achieve conditional generation by pushing the denoising diffusion trajectories away from decision boundaries.<n>We propose a generic postprocessing step built upon flow-matching to shrink the gap between the learned distribution for a pretrained denoising diffusion model and the real data distribution.
  arXiv  Detail & Related papers  (2025-03-13T17:59:59Z)
• Theory on Score-Mismatched Diffusion Models and Zero-Shot Conditional Samplers [49.97755400231656]
  We present the first performance guarantee with explicit dimensional dependencies for general score-mismatched diffusion samplers.<n>We show that score mismatches result in an distributional bias between the target and sampling distributions, proportional to the accumulated mismatch between the target and training distributions.<n>This result can be directly applied to zero-shot conditional samplers for any conditional model, irrespective of measurement noise.
  arXiv  Detail & Related papers  (2024-10-17T16:42:12Z)
• Rejection via Learning Density Ratios [50.91522897152437]
  Classification with rejection emerges as a learning paradigm which allows models to abstain from making predictions.<n>We propose a different distributional perspective, where we seek to find an idealized data distribution which maximizes a pretrained model's performance.<n>Our framework is tested empirically over clean and noisy datasets.
  arXiv  Detail & Related papers  (2024-05-29T01:32:17Z)
• 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)
• The Implicit Bias of Batch Normalization in Linear Models and Two-layer Linear Convolutional Neural Networks [117.93273337740442]
  We show that gradient descent converges to a uniform margin classifier on the training data with an $exp(-Omega(log2 t))$ convergence rate. We also show that batch normalization has an implicit bias towards a patch-wise uniform margin.
  arXiv  Detail & Related papers  (2023-06-20T16:58:00Z)
• Variational Classification [51.2541371924591]
  We derive a variational objective to train the model, analogous to the evidence lower bound (ELBO) used to train variational auto-encoders. Treating inputs to the softmax layer as samples of a latent variable, our abstracted perspective reveals a potential inconsistency. We induce a chosen latent distribution, instead of the implicit assumption found in a standard softmax layer.
  arXiv  Detail & Related papers  (2023-05-17T17:47:19Z)
• Classifier-Free Diffusion Guidance [17.355749359987648]
  guidance is a recently introduced method of trade off mode coverage and sample fidelity in conditional diffusion models. We show that guidance can be indeed performed by a pure generative model without such a classifier. We combine the resulting conditional and unconditional score estimates to attain a trade-off between sample quality and diversity.
  arXiv  Detail & Related papers  (2022-07-26T01:42:07Z)
• 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)
• Approximate Bayesian Computation via Classification [0.966840768820136]
  Approximate Computation (ABC) enables statistical inference in complex models whose likelihoods are difficult to calculate but easy to simulate from. ABC constructs a kernel-type approximation to the posterior distribution through an accept/reject mechanism which compares summary statistics of real and simulated data. We consider the traditional accept/reject kernel as well as an exponential weighting scheme which does not require the ABC acceptance threshold.
  arXiv  Detail & Related papers  (2021-11-22T20:07:55Z)
• Scalable Cross Validation Losses for Gaussian Process Models [22.204619587725208]
  We use Polya-Gamma auxiliary variables and variational inference to accommodate binary and multi-class classification. We find that our method offers fast training and excellent predictive performance.
  arXiv  Detail & Related papers  (2021-05-24T21:01:47Z)
• Understanding Classifier Mistakes with Generative Models [88.20470690631372]
  Deep neural networks are effective on supervised learning tasks, but have been shown to be brittle. In this paper, we leverage generative models to identify and characterize instances where classifiers fail to generalize. Our approach is agnostic to class labels from the training set which makes it applicable to models trained in a semi-supervised way.
  arXiv  Detail & Related papers  (2020-10-05T22:13:21Z)
• 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)
• Consistency Regularization for Certified Robustness of Smoothed Classifiers [89.72878906950208]
  A recent technique of randomized smoothing has shown that the worst-case $ell$-robustness can be transformed into the average-case robustness. We found that the trade-off between accuracy and certified robustness of smoothed classifiers can be greatly controlled by simply regularizing the prediction consistency over noise.
  arXiv  Detail & Related papers  (2020-06-07T06:57:43Z)
• Self-adaptive Re-weighted Adversarial Domain Adaptation [12.73753413032972]
  We present a self-adaptive re-weighted adversarial domain adaptation approach. It tries to enhance domain alignment from the perspective of conditional distribution. Empirical evidence demonstrates that the proposed model outperforms state of the arts on standard domain adaptation datasets.
  arXiv  Detail & Related papers  (2020-05-30T08:35:18Z)

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.