Pith. sign in

REVIEW 9 cited by

Fractal Generative Models

Not yet reviewed by Pith; the record is open.

Re-run · record.json Download PDF Read on arXiv ↗

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2502.17437 v2 pith:5MS3QVQE submitted 2025-02-24 cs.LG cs.CV

Tianhong Li , Qinyi Sun , Lijie Fan , Kaiming He This is my paper
classification cs.LGcs.CV
keywords generativeatomicfractalmodelsmodulesabstractinggenerationmodularization
verification ladder T0 review T1 audit T2 compute T3 formal
0 comments
read the original abstract

Modularization is a cornerstone of computer science, abstracting complex functions into atomic building blocks. In this paper, we introduce a new level of modularization by abstracting generative models into atomic generative modules. Analogous to fractals in mathematics, our method constructs a new type of generative model by recursively invoking atomic generative modules, resulting in self-similar fractal architectures that we call fractal generative models. As a running example, we instantiate our fractal framework using autoregressive models as the atomic generative modules and examine it on the challenging task of pixel-by-pixel image generation, demonstrating strong performance in both likelihood estimation and generation quality. We hope this work could open a new paradigm in generative modeling and provide a fertile ground for future research. Code is available at https://github.com/LTH14/fractalgen.

Discussion (0). Sign in to comment.

Forward citations

Cited by 9 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. A Frozen Pixel-Space Diffusion Model Can Guide Itself with Its Own Samples

    cs.CV 2026-07 conditional novelty 7.0 of 10

    Training a small adapter on a frozen pixel diffusion model's own samples and extrapolating the final prediction away from the adapter's intermediate prediction improves FID on ImageNet.

  2. Revisiting Autoregressive Models for Generative Image Classification

    cs.CV 2026-03 accept novelty 6.5 of 10

    Order-marginalized any-order AR models (RandAR) outperform diffusion generative classifiers on ImageNet and OOD sets and match strong SSL models at far lower cost.

  3. DuSPiT: Dual-Branch Sub-Patch Pixel Diffusion Transformer

    cs.CV 2026-07 conditional novelty 6.0 of 10

    DuSPiT splits pixel diffusion into a compact global-structure branch and a high-capacity subpatch detail branch, reaching ImageNet-512 FID 1.52 at 329 GFLOPs, below JiT-G/32 (1.78).

  4. Signal or Noise? Understanding Generative Models for Real-World Sensor Time Series

    cs.LG 2026-07 conditional novelty 6.0 of 10

    Across 14 sensor generation settings, flow-matching models are the strongest overall baseline, while demographic conditioning, time-frequency modeling, and moderate synthetic augmentation improve hard regimes and down...

  5. PixNerd: Pixel Neural Field Diffusion

    cs.CV 2025-07 conditional novelty 6.0 of 10

    PixNerd is a single-stage pixel-space diffusion transformer that uses predicted neural field weights to decode large patches, reaching 2.15 FID on ImageNet 256 without a VAE.

  6. STARFlow: Scaling Latent Normalizing Flows for High-resolution Image Synthesis

    cs.CV 2025-06 conditional novelty 6.0 of 10

    A latent-space transformer autoregressive flow with one deep block plus shallow refiners, tuned noise injection, and score-based guidance reaches competitive FID in high-resolution image synthesis, the first at this s...

  7. Ctrl-Z Sampling: Scaling Diffusion Sampling with Controlled Random Zigzag Explorations

    cs.CV 2025-06 conditional novelty 5.0 of 10

    Ctrl-Z Sampling improves text-to-image outputs by adaptively rolling back and re-exploring when a reward model flags a quality plateau, at roughly 3 to 9 times the usual compute.

  8. FRN: Fractal-Based Recursive Spectral Reconstruction Network

    cs.CV 2025-05 conditional novelty 5.0 of 10

    A recursive network that builds hyperspectral bands progressively from RGB using a shared atomic module reports state-of-the-art reconstruction on CAVE and Harvard with only 0.30M parameters.

  9. Fractal Flow: Hierarchical and Interpretable Normalizing Flow via Topic Modeling and Recursive Strategy

    stat.ML 2025-08 reject novelty 4.0 of 10

    Fractal Flow combines a Dirichlet-topic latent prior with recursive coupling layers in a normalizing flow, reporting lower bits-per-dim than a custom RealNVP baseline on MNIST and FashionMNIST.

Pith tools