Research map / Learning theory and algorithms / Prediction and decision losses
Don't Forget Its Variance! The Minimum Path Variance Principle for Accurate and Stable Score-Based Models
Wei Chen, Jiacheng Li, Shigui Li, Zhiqi Lin, Junmei Yang, John Paisley, Delu Zeng
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 |
|---|---|---|---|
| use learned minimum-path-variance schedule (MVP affine) instead of Linear path | density ratio estimation · MI estimation d=160 (MI=40) | MSE: 9.17 -> 1.02 | Improved |
| use MVP (spherical) instead of VP path | density estimation · BSDS300 | test NLL: -131.90 -> -143.97 | Improved |
| use MVP (affine) instead of VP path | density estimation · POWER | test NLL: -0.04 -> -0.81 | Improved |
| use MVP (affine) instead of VP path | density estimation · GAS | test NLL: -6.02 -> -6.95 | Improved |
| use MVP (spherical) instead of VP path | density estimation · MINIBOONE | test NLL: 18.25 -> 17.81 | Improved |
| use MVP (spherical) instead of VP path | density estimation · HEPMASS | test NLL: 18.06 -> 18.01 | Improved |
| increase KMM components from 1 to 2 (ablation) | MVP path · MI estimation d=160 | MSE: 1.32 -> 1.09 | Improved |
| increase KMM components from 1 to 2 (ablation) | MVP path · BSDS300 | test NLL: -48.51 -> -145.44 | Improved |
| increase KMM components from 2 to 5 (ablation) | MVP path · MI estimation d=160 | MSE: 1.09 -> 1.02 | Improved |
| increase KMM components from 5 to 8 (ablation) | MVP path · MI estimation d=160 | MSE: 1.02 -> 2.32 | Hurt |
| increase KMM components from 5 to 8 (ablation) | MVP path · BSDS300 | test NLL: -143.97 -> -143.60 | No clear effect |
| choose spherical instead of affine constraint (ablation) | MVP path · BSDS300 | test NLL: -129.73 -> -143.97 | Improved |
Closest papers by meaning
- Neural Network-Based Score Estimation in Diffusion Models: Optimization and Generalization (ICLR 2024 · similarity 0.85)
- The Bias-Variance Tradeoff in Data-Driven Optimization: A Local Misspecification Perspective (NeurIPS 2025 · similarity 0.85)
- Minimal Variance Model Aggregation: A principled, non-intrusive, and versatile integration of black box models (ICLR 2025 · similarity 0.84)
- Score-based generative models are provably robust: an uncertainty quantification perspective (NeurIPS 2024 · similarity 0.84)
- Tuning Stable Rank Shrinkage: Aiming at the Overlooked Structural Risk in Fine-tuning (CVPR 2024 · similarity 0.83)
- On the Variance of Neural Network Training with respect to Test Sets and Distributions (ICLR 2024 · similarity 0.83)
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.
