Seasoning Generative Models for a Generalization Aftertaste

• URL: http://arxiv.org/abs/2603.18817v1
• Date: Thu, 19 Mar 2026 12:13:03 GMT
• Title: Seasoning Generative Models for a Generalization Aftertaste
• Abstract summary: We show that refined generative models provably improve generalization, compared to its non-refined counterpart.<n>Our work provides a theoretical validation for existing work, suggests avenues for new algorithms, and contributes to our understanding of generalization in generative models at large.
• Score: 32.99427367936986
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: The use of discriminators to train or fine-tune generative models has proven to be a rather successful framework. A notable example is Generative Adversarial Networks (GANs) that minimize a loss incurred by training discriminators along with other paradigms that boost generative models via discriminators that satisfy weak learner constraints. More recently, even diffusion models have shown advantages with some kind of discriminator guidance. In this work, we extend a strong-duality result related to $f$-divergences which gives rise to a discriminator-guided recipe that allows us to \textit{refine} any generative model. We then show that the refined generative models provably improve generalization, compared to its non-refined counterpart. In particular, our analysis reveals that the gap in generalization is improved based on the Rademacher complexity of the discriminator set used for refinement. Our recipe subsumes a recently introduced score-based diffusion approach (Kim et al., 2022) that has shown great empirical success, however allows us to shed light on the generalization guarantees of this method by virtue of our analysis. Thus, our work provides a theoretical validation for existing work, suggests avenues for new algorithms, and contributes to our understanding of generalization in generative models at large.

Related papers

• Revisiting Autoregressive Models for Generative Image Classification [24.39392782908908]
  We revisit visual AR-based generative classifiers and identify an important limitation of prior approaches.<n>We observe that single-order predictions rely more on partial discriminative cues, while averaging over multiple token orders provides a more comprehensive signal.<n>We leverage recent any-order AR models to estimate order-marginalized predictions, unlocking the high classification potential of AR models.
  arXiv  Detail & Related papers  (2026-03-19T16:44:23Z)
• Mitigating Reward Hacking in RLHF via Bayesian Non-negative Reward Modeling [49.41422138354821]
  We propose a principled reward modeling framework that integrates non-negative factor analysis into the Bradley-Terry preference model.<n>BNRM represents rewards through a sparse, non-negative latent factor generative process.<n>We show that BNRM substantially mitigates reward over-optimization, improves robustness under distribution shifts, and yields more interpretable reward decompositions than strong baselines.
  arXiv  Detail & Related papers  (2026-02-11T08:14:11Z)
• Generative Classifiers Avoid Shortcut Solutions [84.23247217037134]
  Discriminative approaches to classification often learn shortcuts that hold in-distribution but fail under minor distribution shift.<n>We show that generative classifiers can avoid this issue by modeling all features, both core and spurious, instead of mainly spurious ones.<n>We find that diffusion-based and autorerimigressive generative classifiers achieve state-of-the-art performance on five standard image and text distribution shift benchmarks.
  arXiv  Detail & Related papers  (2025-12-31T18:31:46Z)
• Did Models Sufficient Learn? Attribution-Guided Training via Subset-Selected Counterfactual Augmentation [61.248535801314375]
  Subset-Selected Counterfactual Augmentation (SS-CA)<n>We develop Counterfactual LIMA to identify minimal spatial region sets whose removal can selectively alter model predictions.<n>Experiments show that SS-CA improves generalization on in-distribution (ID) test data and achieves superior performance on out-of-distribution (OOD) benchmarks.
  arXiv  Detail & Related papers  (2025-11-15T08:39:22Z)
• Constrained Auto-Regressive Decoding Constrains Generative Retrieval [71.71161220261655]
  Generative retrieval seeks to replace traditional search index data structures with a single large-scale neural network.<n>In this paper, we examine the inherent limitations of constrained auto-regressive generation from two essential perspectives: constraints and beam search.
  arXiv  Detail & Related papers  (2025-04-14T06:54:49Z)
• Improving Discriminator Guidance in Diffusion Models [43.91753296748528]
  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.
  arXiv  Detail & Related papers  (2025-03-20T13:04:43Z)
• Precision-Recall Divergence Optimization for Generative Modeling with GANs and Normalizing Flows [54.050498411883495]
  We develop a novel training method for generative models, such as Generative Adversarial Networks and Normalizing Flows. We show that achieving a specified precision-recall trade-off corresponds to minimizing a unique $f$-divergence from a family we call the textitPR-divergences. Our approach improves the performance of existing state-of-the-art models like BigGAN in terms of either precision or recall when tested on datasets such as ImageNet.
  arXiv  Detail & Related papers  (2023-05-30T10:07:17Z)
• Enhancing Text Generation with Cooperative Training [23.971227375706327]
  Most prevailing methods trained generative and discriminative models in isolation, which left them unable to adapt to changes in each other. We introduce a textitself-consistent learning framework in the text field that involves training a discriminator and generator cooperatively in a closed-loop manner. Our framework are able to mitigate training instabilities such as mode collapse and non-convergence.
  arXiv  Detail & Related papers  (2023-03-16T04:21:19Z)
• Understanding Why Generalized Reweighting Does Not Improve Over ERM [36.69039005731499]
  Empirical risk minimization (ERM) is known in practice to be non-robust to distributional shift where the training and the test distributions are different. A suite of approaches, such as importance weighting, and variants of distributionally robust optimization (DRO) have been proposed to solve this problem. But a line of recent work has empirically shown that these approaches do not significantly improve over ERM in real applications with distribution shift.
  arXiv  Detail & Related papers  (2022-01-28T17:58:38Z)
• xFAIR: Better Fairness via Model-based Rebalancing of Protected Attributes [15.525314212209564]
  Machine learning software can generate models that inappropriately discriminate against specific protected social groups. We propose xFAIR, a model-based extrapolation method, that is capable of both mitigating bias and explaining the cause.
  arXiv  Detail & Related papers  (2021-10-03T22:10:14Z)
• Generalization Properties of Optimal Transport GANs with Latent Distribution Learning [52.25145141639159]
  We study how the interplay between the latent distribution and the complexity of the pushforward map affects performance. Motivated by our analysis, we advocate learning the latent distribution as well as the pushforward map within the GAN paradigm.
  arXiv  Detail & Related papers  (2020-07-29T07:31:33Z)
• Mind the Trade-off: Debiasing NLU Models without Degrading the In-distribution Performance [70.31427277842239]
  We introduce a novel debiasing method called confidence regularization. It discourages models from exploiting biases while enabling them to receive enough incentive to learn from all the training examples. We evaluate our method on three NLU tasks and show that, in contrast to its predecessors, it improves the performance on out-of-distribution datasets.
  arXiv  Detail & Related papers  (2020-05-01T11:22:55Z)

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.