⌁ Awesome World Model Hub Papers · Datasets · Projects
中文 GitHub
← Back to papers

AlayaVista: Streaming World Modeling from Panoramic States to Perspective Video

arXiv 2026 46.3 method

TLDR

AlayaVista is a camera-controllable streaming video world model that decouples panoramic latent scene evolution from perspective viewport rendering, trained with a large-scale real-world panoramic video dataset.

Reasoning

Strengths: novel architecture decoupling global panorama and local viewport for efficient streaming; introduces large-scale real-world dataset MUGEN. Weaknesses: abstract lacks quantitative results or comparisons; limited evidence of interactive or RL applicability.

Read-first score

Read-first score 46.3, weighted from topical fit, citation, graph, method, reproducibility, and recency signals. Original total remains 52.

Methodology quality 15%
100

Screens visible abstract and analysis fields for experiment, dataset, baseline, metric, and limitation evidence. markers=benchmark,dataset,experiment,metric,result

Recency 5%
100

Uses a gentle age decay so recent papers surface without erasing older foundations. 2026

Topical relevance 25%
74.3

Uses existing LLM keyword relevance scores normalized to 0-100. world model,world simulator,generative world model,interactive world model,video world model,world dynamics prediction,model-based reinforcement learning world model

Reproducibility 15%
46

Screens links and visible text for paper, code, dataset, artifact, and repository signals. pdf=True; code=False; dataset=False; markers=artifact,dataset

Graph prestige 15%
5.7

PageRank-style prestige over the local citation graph built from available OpenAlex references. local_pagerank=0.0016

Citation impact 15%
0

Uses OpenAlex-shaped citation metadata as a bibliometric attention signal, separate from paper quality. cited_by_count=0

Citation velocity 10%
0

Citation velocity estimates citations per publication-year to reduce old-paper bias. velocity=0.00

Field roles

FrontierMethodology anchor

Rank sensitivity

Stability: volatile; rank range: 834.

Keyword Scores

world model
10
video world model
10
generative world model
9
interactive world model
8
world dynamics prediction
8
world simulator
7
model-based reinforcement learning world model
0

Deep Analysis

Innovations

  • Decouples panoramic world evolution from perspective observation synthesis in a streaming video world model.
  • Constructs a 360-degree scene prior from a single perspective image using a pretrained panorama expansion model.
  • Evolves scenes as camera-conditioned panoramic latent states and renders requested perspective video latents through a latent viewport renderer.
  • Adds a perspective refiner for detail restoration, artifact suppression, and super-resolution.
  • Adapts panoramic generation to chunk-autoregressive generation and distills panoramic generation and perspective refinement into few-step processes for efficient streaming.
  • Introduces MUGEN, a large-scale real-world panoramic video dataset with 1,318 hours of at least 4K video and rich semantic and geometric annotations.

Methodology

AlayaVista begins with a single perspective image, expands it into a 360-degree scene prior via a pretrained panorama expansion model, and evolves the scene as a camera-conditioned panoramic latent state. A latent viewport renderer maps this state to perspective video latents, while a perspective refiner restores details, suppresses artifacts, and performs super-resolution; the panoramic generator uses chunk-autoregressive generation and both panoramic generation and perspective refinement are distilled into few-step processes.

Key Results

The abstract does not report quantitative experimental results or benchmark metrics; it highlights the MUGEN dataset scale of 1,318 hours of real-world panoramic video at resolutions of at least 4K.

Limitations

  • No limitations are explicitly stated in the abstract.

Resources

Paper

Tags