Visionary: The World Model Carrier Built on WebGPU-Powered Gaussian Splatting Platform

arXiv 25.12 2025 35.1 system

TLDR

Visionary is a web-native platform for real-time Gaussian Splatting rendering with WebGPU and ONNX inference, enabling dynamic neural processing and generative post-processing.

Reasoning

The paper presents a practical, open-source platform that improves rendering efficiency and lowers deployment barriers, but its contribution is primarily a rendering/inference engine rather than a full world model or simulator. The abstract mentions world models only as motivation, and lacks evidence for world dynamics prediction or model-based RL.

Read-first score

Read-first score 35.1, weighted from topical fit, citation, graph, method, reproducibility, and recency signals. Original total remains 9.

Recency 8%

86.7

Uses a gentle age decay so recent papers surface without erasing older foundations. 2025

Methodology quality 25%

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

Reproducibility 25%

Screens links and visible text for paper, code, dataset, artifact, and repository signals. pdf=True; code=False; dataset=False; markers=none

Topical relevance 42%

12.9

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

Field roles

Frontier

Rank sensitivity

Stability: volatile; rank range: 172.

Keyword Scores

world model

generative world model

interactive world model

world simulator

video world model

world dynamics prediction

model-based reinforcement learning world model

Deep Analysis

Innovations

Open, web-native platform for real-time Gaussian Splatting and meshes rendering using WebGPU
Per-frame ONNX inference enabling dynamic neural processing in the browser
Standardized Gaussian Generator contract for plug-and-play algorithm integration
Feedforward generative post-processing via inference
three.js plugin with concise TypeScript API for seamless web integration
GPU-based primitive sorting achieving superior rendering efficiency over current Web viewers

Methodology

Visionary is built on a WebGPU renderer with per-frame ONNX inference, allowing dynamic neural processing in a lightweight browser environment. It introduces a Gaussian Generator contract that supports standard 3DGS rendering and plug-and-play algorithms to generate or update Gaussians each frame. The platform also offers a three.js library with a TypeScript API for integration into existing web applications.

Key Results

Under identical 3DGS assets, Visionary achieves superior rendering efficiency compared to current Web viewers due to GPU-based primitive sorting. It already supports multiple variants including MLP-based 3DGS, 4DGS, neural avatars, and style transformation or enhancement networks.

Resources

Paper