AI Paper Atlas

Research map / Image generation / Fast diffusion sampling

Solving Linear Inverse Problems Provably via Posterior Sampling with Latent Diffusion Models

Litu Rout, Negin Raoof, Giannis Daras, Constantine Caramanis, Alex Dimakis, Sanjay Shakkottai

NeurIPS 2023 · Image generation › Fast diffusion sampling · approach: Diffusion posterior sampling · 5 citations (OpenAlex)

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 changedWhereMeasuredEffect
PSLD with Stable Diffusion instead of DPS on random inpaintingPSLD · FFHQ 256 random inpaintingFID: 33.48 -> 21.34Improved
PSLD instead of DPS on box inpaintingPSLD · FFHQ 256 box inpaintingFID: 35.14 -> 43.11Hurt
PSLD instead of DPS on box inpaintingPSLD · FFHQ 256 box inpaintingLPIPS: 0.216 -> 0.167Improved
PSLD instead of DPS on 4x super-resolutionPSLD · FFHQ 256 SR 4xFID: 39.35 -> 34.28Improved
PSLD instead of DPS on Gaussian deblurPSLD · FFHQ 256 Gaussian deblurFID: 44.05 -> 41.53Improved
PSLD instead of DPS on super-resolution at 2x, 3x, 4xPSLD · FFHQ 256 SRLPIPS: 0.220 -> 0.185 (2x); 0.247 -> 0.220 (3x); 0.291 -> 0.233 (4x)Improved
Use latent diffusion LDM-VQ-4 (same training data as DPS) for box inpaintingPSLD · FFHQ 256 box inpaintingPSNR: 23.39 -> 24.22Improved
Use LDM-VQ-4 for random inpainting versus DPSPSLD · FFHQ 256 random inpaintingLPIPS: 0.212 -> 0.221No clear effect
Use LDM-VQ-4 for random inpainting versus DPSPSLD · FFHQ 256 random inpaintingPSNR: 29.49 -> 30.31Improved
Use Stable Diffusion instead of LDM-VQ-4 as the foundation modelPSLD · FFHQ 256 random inpaintingLPIPS: 0.221 -> 0.096Improved
Random inpainting with added Gaussian noise sigma=0.05 (PSLD vs DPS)PSLD · FFHQ 256 inpaint + denoisePSNR: 29.49 -> 33.71Improved
Noiseless random inpainting sigma=0.00 (PSLD vs DPS)PSLD · FFHQ 256 inpaint + denoiseLPIPS: 0.171 -> 0.083Improved

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy