Research map / Probabilistic and scientific modeling / Uncertainty and Gaussian processes
Novel Uncertainty Quantification Through Perturbation-Assisted Sample Synthesis
Yifei Liu, Rex Shen, Xiaotong Shen
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 |
|---|---|---|---|
| increase the inference sample size n from 2,050 to 10,000 | PASS-DDPM null FID · CIFAR-10 | FID: 49.75 -> 32.58 | Improved |
| use DCGAN instead of DDPM as candidate generator | PAI image test · CIFAR-10 (n = 10,000) | FID: 32.57 -> 76.17 | Hurt |
| use GLOW instead of DDPM as candidate generator | PAI image test · CIFAR-10 (n = 10,000) | FID: 32.57 -> 61.01 | Hurt |
| add perturbation of size tau up to 1.0 to PASS (ablation) | PASS-TabDDPM · Prediction interval simulation | FID: 0.024 -> 0.024 | No clear effect |
| add perturbation of size tau up to 1.0 to PASS (ablation) | PASS-TabDDPM · Prediction interval simulation | 2-Wasserstein: 1.298 -> 1.298 | No clear effect |
| use PAI intervals instead of conformal inference intervals | PAI prediction intervals · Simulation (alpha = 0.05) | interval length: about 68% of intervals shorter | Improved |
| use PAI instead of conformal inference for coverage | PAI prediction intervals · Simulation (alpha = 0.05) | coverage probability: median 0.95, average 0.9 | Mixed |
| use PASS null distribution instead of asymptotic Gaussian for sentiment test | PAI sentiment test · IMDB | KS test P-value: P-value .004, .002, .000 | Improved |
| compare identical prompts versus different prompts | PAI coherence test · Stable Diffusion prompts | FID score: .544 -> 14.250 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- GLOW — Glow: Generative Flow with Invertible 1x1 Convolutions (NeurIPS 2018) · better on 3 of 3 shared results
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
- Conformal Prediction Beyond the Seen: A Missing Mass Perspective for Uncertainty Quantification in Generative Models (NeurIPS 2025 · similarity 0.85)
- Uncertainty Quantification with the Empirical Neural Tangent Kernel (NeurIPS 2025 · similarity 0.85)
- Efficient Analytic Uncertainty Quantification for Multimodal Regression (NeurIPS 2026 · similarity 0.85)
- Conformal Prediction with Paraphrase-Aware Scoring for LLM Uncertainty Quantification (NeurIPS 2026 · similarity 0.84)
- Uncertainty Quantification for Large Language Diffusion Models (NeurIPS 2026 · similarity 0.84)
- A Bias-Variance-Covariance Decomposition of Kernel Scores for Generative Models (ICML 2024 · similarity 0.84)
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.
