Research map / Image generation / Fast diffusion sampling
Maximum Likelihood Training of Implicit Nonlinear Diffusion Model
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| pre-train the score network with DDPM++ for 500k steps before INDM training (ablation) | INDM (NLL) · CIFAR-10 | FID (ODE): 8.49 -> 6.01 | Helped |
| train INDM from scratch vs DDPM++ baseline (ablation) | INDM (NLL) · CIFAR-10 | FID (ODE): 6.70 -> 8.49 | Hurt |
| pre-training does not change NLL (ablation) | INDM (NLL) · CIFAR-10 | NLL: 2.98 -> 2.98 | No clear effect |
| add nonlinear flow diffusion (INDM) to DDPM++ with likelihood weighting | DDPM++ (VP, NLL) · CIFAR-10 | NELBO: 3.13 -> 2.98 | Helped |
| add INDM flow to DDPM++ with likelihood weighting | DDPM++ (VP, NLL) · CIFAR-10 | variational gap: 0.10 -> 0.00 | Helped |
| add INDM flow to DDPM++ with likelihood weighting | DDPM++ (VP, NLL) · CIFAR-10 | NLL: 3.03 -> 2.98 | Helped |
| add INDM flow to NCSN++ (VE, FID weighting) | NCSN++ (VE) · CIFAR-10 | NLL: 4.86 -> 3.22 | Helped |
| add INDM flow to NCSN++ (VE, FID weighting) | NCSN++ (VE) · CIFAR-10 | FID (PC): 2.38 -> 2.29 | Helped |
| add INDM flow to DDPM++ (VP, FID weighting) | DDPM++ (VP) · CIFAR-10 | FID (ODE): 3.90 -> 3.61 | Helped |
| add INDM flow to DDPM++ (VP, FID weighting) | DDPM++ (VP) · CIFAR-10 | FID (PC): 2.89 -> 2.90 | No clear effect |
| add INDM flow to DDPM++ (VP, NLL weighting) | DDPM++ (VP, NLL) · CIFAR-10 | FID (PC): 5.17 -> 5.30 | Hurt |
| add INDM flow to DDPM++ (VP, FID weighting) | DDPM++ (VP) · CelebA 64x64 | FID: 2.32 -> 1.75 | Helped |
Closest papers by meaning
- One-Step Diffusion Distillation through Score Implicit Matching (NeurIPS 2024 · similarity 0.89)
- Neural Flow Diffusion Models: Learnable Forward Process for Improved Diffusion Modelling (NeurIPS 2024 · similarity 0.88)
- Accelerating Convergence of Score-Based Diffusion Models, Provably (ICML 2024 · similarity 0.88)
- Understanding Generalizability of Diffusion Models Requires Rethinking the Hidden Gaussian Structure (NeurIPS 2024 · similarity 0.88)
- What’s the score? Automated Denoising Score Matching for Nonlinear Diffusions (ICML 2024 · similarity 0.87)
- Nonparametric Estimation of a Factorizable Density using Diffusion Models (NeurIPS 2026 · similarity 0.87)
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.