Salvato in:
| Autori principali: | , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | https://arxiv.org/abs/2510.02826 |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866917275153465344 |
|---|---|
| author | Hong, Steve Belkadi, Samuel |
| author_facet | Hong, Steve Belkadi, Samuel |
| contents | We reinterpret Visual Autoregressive (VAR) models as iterative refinement models to identify which design choices drive their quality-efficiency trade-off. Instead of treating VAR only as next-scale autoregression, we formalise it as a deterministic forward process that builds a Laplacian-style latent pyramid, together with a learned backward process that reconstructs samples in a small number of coarse-to-fine steps. This formulation makes the link to denoising diffusion explicit and highlights three modelling choices that may underlie VAR's efficiency and sample quality: refinement in a learned latent space, discrete prediction over code indices, and decomposition by spatial frequency. We support this view with controlled experiments that isolate the contribution of each factor to quality and speed. We also discuss how the same framework can be adapted to permutation-invariant graph generation and probabilistic medium-range weather forecasting, and how it provides practical points of contact with diffusion methods while preserving few-step, scale-parallel generation. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2510_02826 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Multi-scale Autoregressive Models are Laplacian, Discrete, and Latent Diffusion Models in Disguise Hong, Steve Belkadi, Samuel Machine Learning We reinterpret Visual Autoregressive (VAR) models as iterative refinement models to identify which design choices drive their quality-efficiency trade-off. Instead of treating VAR only as next-scale autoregression, we formalise it as a deterministic forward process that builds a Laplacian-style latent pyramid, together with a learned backward process that reconstructs samples in a small number of coarse-to-fine steps. This formulation makes the link to denoising diffusion explicit and highlights three modelling choices that may underlie VAR's efficiency and sample quality: refinement in a learned latent space, discrete prediction over code indices, and decomposition by spatial frequency. We support this view with controlled experiments that isolate the contribution of each factor to quality and speed. We also discuss how the same framework can be adapted to permutation-invariant graph generation and probabilistic medium-range weather forecasting, and how it provides practical points of contact with diffusion methods while preserving few-step, scale-parallel generation. |
| title | Multi-scale Autoregressive Models are Laplacian, Discrete, and Latent Diffusion Models in Disguise |
| topic | Machine Learning |
| url | https://arxiv.org/abs/2510.02826 |