Improving Discriminator Guidance in Diffusion Models

• URL: http://arxiv.org/abs/2503.16117v1
• Date: Thu, 20 Mar 2025 13:04:43 GMT
• Title: Improving Discriminator Guidance in Diffusion Models
• Authors: Alexandre Verine, Mehdi Inane, Florian Le Bronnec, Benjamin Negrevergne, Yann Chevaleyre,
• Abstract summary: We show that training the discriminator using Cross-Entropy loss, as commonly done, can increase the Kullback-Leibler divergence between the model and target distributions.<n>We propose a theoretically sound training objective for discriminator guidance that properly minimizes the KL divergence.
• Score: 43.91753296748528
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Discriminator Guidance has become a popular method for efficiently refining pre-trained Score-Matching Diffusion models. However, in this paper, we demonstrate that the standard implementation of this technique does not necessarily lead to a distribution closer to the real data distribution. Specifically, we show that training the discriminator using Cross-Entropy loss, as commonly done, can in fact increase the Kullback-Leibler divergence between the model and target distributions, particularly when the discriminator overfits. To address this, we propose a theoretically sound training objective for discriminator guidance that properly minimizes the KL divergence. We analyze its properties and demonstrate empirically across multiple datasets that our proposed method consistently improves over the conventional method by producing samples of higher quality.

Related papers

• DIVE: Inverting Conditional Diffusion Models for Discriminative Tasks [79.50756148780928]
  This paper studies the problem of leveraging pretrained diffusion models for performing discriminative tasks. We extend the discriminative capability of pretrained frozen generative diffusion models from the classification task to the more complex object detection task, by "inverting" a pretrained layout-to-image diffusion model.
  arXiv  Detail & Related papers  (2025-04-24T05:13:27Z)
• Learned Reference-based Diffusion Sampling for multi-modal distributions [2.1383136715042417]
  We introduce Learned Reference-based Diffusion Sampler (LRDS), a methodology specifically designed to leverage prior knowledge on the location of the target modes. LRDS proceeds in two steps by learning a reference diffusion model on samples located in high-density space regions. We experimentally demonstrate that LRDS best exploits prior knowledge on the target distribution compared to competing algorithms on a variety of challenging distributions.
  arXiv  Detail & Related papers  (2024-10-25T10:23:34Z)
• Diffusion Attribution Score: Evaluating Training Data Influence in Diffusion Model [22.39558434131574]
  Existing data attribution methods for diffusion models typically quantify the contribution of a training sample.<n>We argue that the direct usage of diffusion loss cannot represent such a contribution accurately due to the calculation of diffusion loss.<n>We propose Diffusion Attribution Score (textitDAS) to measure the direct comparison between predicted distributions with an attribution score.
  arXiv  Detail & Related papers  (2024-10-24T10:58:17Z)
• Rejection via Learning Density Ratios [50.91522897152437]
  Classification with rejection emerges as a learning paradigm which allows models to abstain from making predictions. We propose a different distributional perspective, where we seek to find an idealized data distribution which maximizes a pretrained model's performance. Our framework is tested empirically over clean and noisy datasets.
  arXiv  Detail & Related papers  (2024-05-29T01:32:17Z)
• Discriminator Guidance for Autoregressive Diffusion Models [12.139222986297264]
  We introduce discriminator guidance in the setting of Autoregressive Diffusion Models. We derive ways of using a discriminator together with a pretrained generative model in the discrete case.
  arXiv  Detail & Related papers  (2023-10-24T13:14:22Z)
• Supervised Contrastive Learning with Heterogeneous Similarity for Distribution Shifts [3.7819322027528113]
  We propose a new regularization using the supervised contrastive learning to prevent such overfitting and to train models that do not degrade their performance under the distribution shifts. Experiments on benchmark datasets that emulate distribution shifts, including subpopulation shift and domain generalization, demonstrate the advantage of the proposed method.
  arXiv  Detail & Related papers  (2023-04-07T01:45:09Z)
• 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)
• Personalized Trajectory Prediction via Distribution Discrimination [78.69458579657189]
  Trarimiy prediction is confronted with the dilemma to capture the multi-modal nature of future dynamics. We present a distribution discrimination (DisDis) method to predict personalized motion patterns. Our method can be integrated with existing multi-modal predictive models as a plug-and-play module.
  arXiv  Detail & Related papers  (2021-07-29T17:42:12Z)
• Exploring Dropout Discriminator for Domain Adaptation [27.19677042654432]
  Adaptation of a classifier to new domains is one of the challenging problems in machine learning. We propose a curriculum based dropout discriminator that gradually increases the variance of the sample based distribution. An ensemble of discriminators helps the model to learn the data distribution efficiently.
  arXiv  Detail & Related papers  (2021-07-09T06:11:34Z)
• Learning Diverse Representations for Fast Adaptation to Distribution Shift [78.83747601814669]
  We present a method for learning multiple models, incorporating an objective that pressures each to learn a distinct way to solve the task. We demonstrate our framework's ability to facilitate rapid adaptation to distribution shift.
  arXiv  Detail & Related papers  (2020-06-12T12:23:50Z)
• When Relation Networks meet GANs: Relation GANs with Triplet Loss [110.7572918636599]
  Training stability is still a lingering concern of generative adversarial networks (GANs) In this paper, we explore a relation network architecture for the discriminator and design a triplet loss which performs better generalization and stability. Experiments on benchmark datasets show that the proposed relation discriminator and new loss can provide significant improvement on variable vision tasks.
  arXiv  Detail & Related papers  (2020-02-24T11:35:28Z)

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.