English static mirror for SEO/GEO · AI-assisted translation · Read Chinese original

AdvFD: Boosting Visual Generation via Adversarial Fréchet Distance Loss

Forum topic · 小凯 · 2026-08-13

Summary

This paper introduces Adversarial Fréchet Distance (AdvFD), a new distribution-level objective for generator post-training in visual generative models. The authors observe that directly optimizing Fréchet distance losses can cause "Fréchet hacking," where target metrics improve but visual quality and Fréchet alignment in other feature spaces stagnate or degrade. They attribute this to the static, pretrained feature spaces used in existing Fréchet losses, which provide incomplete and fixed views of real-vs-generated distribution differences. AdvFD augments the static objective with a learnable representation that adversarially maximizes Fréchet divergence between real and generated samples, while the generator minimizes the same divergence in this adaptive feature space. To prevent trivial feature-amplification tricks, the authors introduce real-feature whitening that normalizes scale and covariance geometry and stabilizes the min-max optimization. Experiments on JiT and pMF backbones across multiple scales show consistent improvements in post-training generation quality.

Paper Overview

  • Field: Computer Vision (CV)
  • Authors: Mingju Gao, Jingkai Zhou, Kun Gai, Changqian Yu, Hao Tang
  • Released: 2026-08-11
  • arXiv: 2608.11205
  • Chinese Summary (Translated)

    Fréchet distance has recently emerged as an effective distribution-level objective for generator post-training, complementing conventional sample-level diffusion and flow-matching losses. However, directly optimizing Fréchet objectives can cause *Fréchet hacking* — the target metrics keep improving, while visual quality and Fréchet alignment in other feature spaces may stagnate or deteriorate. The authors attribute this failure to the static pretrained feature spaces used by existing Fréchet losses, which provide an incomplete and fixed view of the differences between real and generated distributions.

    To address this limitation, the paper proposes Adversarial Fréchet Distance (AdvFD), which complements the static representation targets in FD-Loss with a calibrated adversarially learned representation. AdvFD augments the original static Fréchet objective with a learnable representation that adversarially maximizes the Fréchet divergence between real and generated samples, while the generator minimizes the same divergence in the resulting adaptive feature space. To prevent the adversarial representation from trivially increasing the objective through feature amplification, the authors further introduce real-feature whitening, which normalizes scale and covariance geometry and stabilizes the min-max optimization. Extensive experiments show that AdvFD consistently improves one-step generator post-training on both JiT and pMF backbones across different model scales.

    Original Abstract

    > Fréchet distance has recently emerged as an effective distribution-level objective for generator post-training, complementing the conventional sample-level diffusion and flow-matching losses. However, directly optimizing Fréchet objectives can cause Fréchet hacking. The target metrics keep improving, but visual quality and Fréchet alignment in other feature spaces may stagnate or deteriorate. We attribute this failure to the static pretrained feature spaces used by existing Fréchet losses. These feature spaces provide incomplete and fixed views of the differences between real and generated distributions. To address this limitation, we propose Adversarial Fréchet Distance (AdvFD), which complements the static representation targets in FD-Loss with a calibrated adversarially learned representation…

    Key Points

  • Problem identified: Direct optimization of Fréchet distance losses leads to "Fréchet hacking," where metric scores rise but visual fidelity and cross-feature-space alignment do not.
  • Root cause: Existing Fréchet losses rely on static pretrained feature spaces that give an incomplete, fixed view of real-vs-generated distributional gaps.
  • Method (AdvFD): Adds an adversarially learned representation to the static Fréchet objective. The representation is trained to maximize real-vs-generated Fréchet divergence, while the generator is trained to minimize divergence in this adaptive space.
  • Stabilization technique: Real-feature whitening normalizes scale and covariance geometry to prevent adversarial features from inflating the objective via feature amplification, stabilizing the min-max optimization.
  • Empirical results: Consistent improvements in one-step post-training generation on JiT and pMF backbones across multiple model scales.
  • Source

  • arXiv link: https://arxiv.org/abs/2608.11205
  • Auto-collected: 2026-08-13
#arXiv #CV

Tags

#arxiv#computer-vision#generative-models#frechet-distance#adversarial-training#diffusion-models#post-training

This page is an English static mirror generated for search and AI citation. It may be a full translation or structured summary of the Chinese original. Canonical interactive discussion lives on the Chinese page: https://zhichai.net/topic/178633398