Discriminative Score Function: Turning Pretrained Models into Functional Generative Priors
cs.CV, cs.LG
Submitted: 2026-10-14
Updated: 2026-10-14
Comments: 20 pages. Accepted at the Fortieth Annual Conference on Neural Information Processing Systems (2026)
Project page: https://junhoo.me/dsf
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Terminology
Sources
- Deconstructing Data Reconstruction: Multiclass, Weight Decay and General Losses
- End-to-End Object Detection with Transformers
- DiffEdit: Diffusion-based semantic image editing with mask guidance
- An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale
- Unlocking Feature Visualization for Deeper Networks with MAgnitude Constrained Optimization
- Your Classifier is Secretly an Energy Based Model and You Should Treat it Like One
- Deep Residual Learning for Image Recognition
- Denoising Diffusion Probabilistic Models
- Do not think about pink elephant!
- Neural Tangent Kernel: Convergence and Generalization in Neural Networks
- Directional convergence and alignment in deep learning
- Zero-shot Quantization: A Comprehensive Survey
- Activation Quantization of Vision Encoders Needs Prefixing Registers
- Temporal Ensembling for Semi-Supervised Learning
- Deep Edge Filter: Return of the Human-Crafted Layer in Deep Learning
- CSF: Black-box Fingerprinting via Compositional Semantics for Text-to-Image Models
- Microsoft COCO: Common Objects in Context
- DINOv2: Learning Robust Visual Features without Supervision
- Denoising Diffusion Implicit Models
- Score-Based Generative Modeling through Stochastic Differential Equations
Related papers
- Loss Knows Best: Detecting Annotation Errors in Videos via Loss Trajectories
- AnchorWeave: World-Consistent Video Generation with Retrieved Local Spatial Memories
- Benchmarking the Robustness of Foundation Models for Mammography under Domain Shift
- MambaX-Net: Dual-Input Mamba-Enhanced Cross-Attention Network for Longitudinal MRI Segmentation
- TeleOCR: Navigating Document Parsing Across Digital and Camera-Captured Documents
- A Survey on Efficient Vision-Language-Action Models