DSF：把预训练判别模型变成生成先验的无训练方法

View PDF HTML (experimental)

Abstract:Discriminative models are trained with task-specific objectives such as classification, detection, or representation learning, yet their practical utility depends on generalizing beyond the training task to unseen samples from the same domain. This suggests that a discriminative model may retain an implicit trace of the data manifold beyond its task labels. We ask whether this trace can be exposed as a generative signal without task-specific guidance. We introduce the Discriminative Score Function (DSF), which extracts an image-space generative field from the loss-gradient geometry of a pretrained discriminative model. DSF provides a training-free update rule that moves images toward model-consistent visual regions without retraining or task-specific targets. We apply DSF to diverse pretrained discriminative models, including ResNet classifiers, DETR detectors, and DINOv2 ViT encoders. DSF generates unconditional images from noise and supports conditional guidance, editing, inpainting, and explanation through the same field. DSF-generated samples also provide effective zero-shot quantization calibration when real images or labels are unavailable, preserving feature geometry better than data-free calibration baselines. These results show that discriminative training leaves a generative functional signal for synthesis and calibration. Project page: this https URL.

Comments:	20 pages. Accepted at the Fortieth Annual Conference on Neural Information Processing Systems (2026)
Subjects:	Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
Cite as:	arXiv:2610.19743 [cs.CV]
	(or arXiv:2610.19743v1 [cs.CV] for this version)
	https://doi.org/10.48550/arXiv.2610.19743 arXiv-issued DOI via DataCite (pending registration)

Submission history

From: Hyeonjin Kim [view email]
[v1] Wed, 14 Oct 2026 08:04:30 UTC (25,606 KB)