Research map / Probabilistic and scientific modeling / Uncertainty and Gaussian processes
CARD: Classification and Regression Diffusion Models
Xizewen Han, Huangjie Zheng, Mingyuan Zhou
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| wrap a pre-trained classifier with CARD diffusion and majority vote | ResNet-18 classifier · CIFAR-10 | accuracy (%): 90.39 -> 90.93 | Improved |
| wrap a pre-trained classifier with CARD diffusion | LeNet classifier · FashionMNIST | accuracy (%): 91.12 -> 91.79 | Improved |
| use CARD diffusion instead of Deep Ensembles | regression · UCI Kin8nm | RMSE: 8.65 -> 6.32 | Improved |
| use CARD diffusion instead of MC Dropout | regression · UCI Wine | RMSE: 0.62 -> 0.63 | No clear effect |
| use CARD diffusion instead of Deep Ensembles | regression · UCI Concrete | NLL: 2.93 -> 2.96 | No clear effect |
| use CARD diffusion instead of PBP/MC Dropout | regression · UCI Wine | QICE (%): 2.79 -> 3.39 | Hurt |
| use CARD diffusion instead of MC Dropout | regression · UCI Naval | QICE (%): 12.50 -> 0.80 | Improved |
| use CARD instead of CM-MF-VI BNN | classification · CIFAR-10 | NLL: 0.39 -> 0.46 | Hurt |
| use CARD instead of MAP | classification · CIFAR-10 | accuracy (%): 84.69 -> 90.93 | Improved |
| pick class by narrowest PIW rather than majority vote (ablation) | CARD · CIFAR-10 | accuracy (%): 90.93 -> 87.84 | Hurt |
Later papers that compared with this method
4 later accepted papers report CARD as a baseline in their results tables; 2 of them report a better number on most of the results they share with it.
- Learning General Causal Structures with Hidden Dynamic Process for Climate Analysis (ICML 2026 · better on 1 of 2 shared results)
- Supervised Score-Based Modeling by Gradient Boosting (AAAI 2025 · better on 6 of 6 shared results)
- Enhancing Uncertainty Estimation and Interpretability with Bayesian Non-negative Decision Layer (ICLR 2025 · better on 6 of 6 shared results)
- On the Identification of Temporal Causal Representation with Instantaneous Dependence (ICLR 2025 · better on 0 of 1 shared result)
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Zero-Shot Conditioning of Score-Based Diffusion Models by Neuro-Symbolic Constraints (AAAI 2025 · similarity 0.87)
- A Likelihood Based Approach to Distribution Regression Using Conditional Deep Generative Models (ICML 2025 · similarity 0.86)
- Label-Noise Robust Diffusion Models (ICLR 2024 · similarity 0.86)
- Training Diffusion-based Generative Models with Limited Data (ICML 2025 · similarity 0.86)
- No Training, No Problem: Rethinking Classifier-Free Guidance for Diffusion Models (ICLR 2025 · similarity 0.85)
- Conditional Diffusion Models are Minimax-Optimal and Manifold-Adaptive for Conditional Distribution Estimation (ICLR 2025 · similarity 0.85)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
