Research map / Image generation / Autoregressive and latent generation
Stabilizing Self-Consuming Diffusion Models with Latent Space Filtering
Zhongteng Cai, Yaxuan Wang, Yang Liu, Xueru Zhang
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 |
|---|---|---|---|
| filter accumulated data by probing-classifier confidence instead of random sampling (ACUR) | ACUR · MNIST | FID | Improved |
| filter accumulated data by probing-classifier confidence instead of random sampling (ablation) | ACUR · MNIST | real image ratio in filtered set | Improved |
| filter by batch-level OLE score instead of random sampling (ablation) | ACUR · MNIST | self-consuming stability | No clear effect |
| add 30% fresh real data per generation (SYN-ADD) to pure synthetic training | SYN · MNIST and CelebA | FID, precision, recall | Improved |
| accumulate historical data (ACU) instead of pure synthetic loop | SYN · MNIST and CelebA | FID, precision, recall | Improved |
| add score-function extrapolation (SIMS) across generations | ACUR · MNIST and CelebA | FID, precision, recall | Hurt |
| add self-correction toward real cluster centers (SC) | ACUR · MNIST and CelebA | recall and FID | Hurt |
| replace full accumulation with LSF on fixed budget (storage) | ACU-LSF · CelebA | training data storage: 582 MB -> 109 MB | Improved |
Closest papers by meaning
- A Closer Look at Model Collapse: From a Generalization-to-Memorization Perspective (NeurIPS 2025 · similarity 0.89)
- Collapse or Thrive: Perils and Promises of Synthetic Data in a Self-Generating World (ICML 2025 · similarity 0.89)
- CORAL: Disentangling Latent Representations in Long-Tailed Diffusion (NeurIPS 2025 · similarity 0.89)
- Towards Theoretical Understandings of Self-Consuming Generative Models (ICML 2024 · similarity 0.88)
- Stabilize the Latent Space for Image Autoregressive Modeling: A Unified Perspective (NeurIPS 2024 · similarity 0.88)
- Probing the Latent Hierarchical Structure of Data via Diffusion Models (ICLR 2025 · similarity 0.88)
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.
