AI Paper Atlas

Research map / Image generation / Autoregressive and latent generation

Unsupervised Representation Learning from Pre-trained Diffusion Probabilistic Models

Zijian Zhang, Zhou Zhao, Zhijie Lin

NeurIPS 2022 · Image generation › Autoregressive and latent generation · approach: Latent diffusion with learned tokenizers · 15 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
adopt PDAE (pre-trained DPM plus mean-shift estimator) instead of Diff-AEautoencoding with inferred x_T · CelebA-HQ reconstructionSSIM: 0.991 -> 0.994Improved
adopt PDAE instead of Diff-AEautoencoding with random x_T · CelebA-HQ reconstructionLPIPS: 0.073 -> 0.098Hurt
adopt PDAE instead of Diff-AEautoencoding with random x_T · CelebA-HQ reconstructionSSIM: 0.677 -> 0.689Improved
reuse pre-trained DPM encoder and redesign loss weighting (PDAE vs Diff-AE)representation learning · training throughput and convergenceimgs/sec/A100: 75.41 -> 81.57Improved
use PDAE instead of Diff-AE for few-shot conditional generationCelebA64 few-shot · Binary MaleFID: 11.52 -> 11.21Improved
use PDAE instead of Diff-AE for few-shot conditional generationCelebA64 few-shot · Binary BlondFID: 16.10 -> 16.96Hurt
use PDAE instead of Diff-AE for few-shot conditional generationCelebA64 PU few-shot · PU BlondFID: 7.01 -> 6.34Improved
use simple loss weighting lambda_t=1 instead of SNR-based redesigned weighting (ablation)PDAE training · training stabilityconvergenceImproved
apply PDAE gradient estimator only in the first 70% of sampling steps for unconditional sampling (ablation)pre-trained DPM sampling · unconditional FFHQ/Horse/Bedroom/CelebA64FIDImproved

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