arXiv cs.AI· Raluca Stevenson, David Bignell, Jennifer C. White, Tabish Rashid, Emanuele Aiello, Mariya Hendriksen, Junha Hyung, Sadegh Aliakbarian, Abdelhak Lemkhenter, Marko Tot, Katja Hofmann, Sarah Parisot·· 56 分钟前AI 评分42
迈向数据高效、实时的世界模型
Towards Data-Efficient, Real-Time World Models
AI 导读
一篇 arXiv 论文提出用两阶段离散扩散方法取代 next-token prediction,并引入轻量级 refinement transformer 进行迭代式 masked 预测。在自回归世界模型基础上实现 79x 推理提速,仅有适度的质量折损;模型仅需约一天(28 小时)游戏数据即可复现高质量、一致的游戏画面生成。
正文 · 原文
Authors:Raluca Stevenson, David Bignell, Jennifer C. White, Tabish Rashid, Emanuele Aiello, Mariya Hendriksen, Junha Hyung, Sadegh Aliakbarian, Abdelhak Lemkhenter, Marko Tot, Katja Hofmann, Sarah Parisot
Abstract:State-of-the-art world models have been used to produce sequences of gameplay that accord with provided user-input actions. However, high quality, consistent gameplay generation often comes at the cost of inference speed, making real-time interactive play and ideation challenging. The substantial quantities of data used to train state-of-the-art models have also made custom training of these models appear impractical, limiting many of their potential uses in creative fields. In this work we demonstrate a major step towards efficient world models: achieving both fast high-quality inference, and improved data efficiency. This is done by replacing the next-token prediction paradigm with a two-stage discrete diffusion-based approach, introducing a lightweight refinement transformer which carries out iterative masked predictions. Building on an autoregressive world model capable of generating highly consistent and complex sequences over minutes (Kanervisto et al., 2025), we achieve substantial model speed-up (79x) with a modest quality trade-off. Additionally, we demonstrate that this model can be trained to reproduce high-quality, consistent gameplay using a dataset containing approximately one day (28 hours) of gameplay data. Our results demonstrate progress towards removing barriers to the use of world models in settings where real-time inference and data efficiency are essential.
| Comments: | 18 pages, 12 figures, 1 table. Includes supplementary material |
| Subjects: | Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2611.22497 [cs.AI] |
| (or arXiv:2611.22497v1 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2611.22497 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Katja Hofmann [view email]
[v1]
Tue, 17 Nov 2026 09:52:15 UTC (8,402 KB)
来源:arXiv cs.AI · arxiv.org