arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2511.02091cs.LGcs.AI
收藏

结构化世界模型的自然构建模块:理论、证据与扩展

Natural Building Blocks for Structured World Models: Theory, Evidence, and Scaling

VERSES AI Research Lab(VERSES AI研究实验室);University of Tübingen(图宾根大学);ELLIS Institute, Tübingen(图宾根ELLIS研究所);MPI for Intelligent Systems, Tübingen(图宾根智能系统研究所)

Lancelot Da Costa, Sanjeev Namjoshi, Mohammed Abbas Ansari, Bernhard Schölkopf

AI总结:

提出基于离散与连续随机过程分层组合的自然构建模块框架,通过固定因果架构并搜索四个深度参数,在保持可解释性的同时实现与神经方法相当的被动建模与主动控制性能。

摘要:

世界建模领域目前处于碎片化状态,研究人员开发的定制架构很少能相互借鉴。我们提出了一个框架,基于任何世界模型都必须捕获的基本随机过程(离散过程(逻辑、符号)和连续过程(物理、动力学)),为结构化世界模型指定了自然构建模块;世界模型随后由这些构建模块的分层组合来定义。我们将隐马尔可夫模型(HMMs)和切换线性动力系统(sLDS)作为离散和连续建模的自然构建模块进行考察——当引入动作扩展时,它们便成为部分可观测马尔可夫决策过程(POMDPs)和受控sLDS。这种模块化方法在同一架构内同时支持被动建模(生成、预测)和主动控制(规划、决策)。我们通过基本固定因果架构并仅搜索四个深度参数,避免了传统结构学习的组合爆炸。我们通过多模态生成建模(被动)和基于像素的规划(主动)来回顾其实际表达能力,其性能与神经方法相当,同时保持了可解释性。核心未决挑战是可扩展的联合结构-参数学习;当前方法通过巧妙地增量增长结构和参数来规避此问题,但在可扩展性方面受限。若能解决,这些自然构建模块可为世界建模提供基础架构,类似于标准化层如何推动了深度学习的进展。

英文摘要:

The field of world modeling is fragmented, with researchers developing bespoke architectures that rarely build upon each other. We propose a framework that specifies the natural building blocks for structured world models based on the fundamental stochastic processes that any world model must capture: discrete processes (logic, symbols) and continuous processes (physics, dynamics); the world model is then defined by the hierarchical composition of these building blocks. We examine Hidden Markov Models (HMMs) and switching linear dynamical systems (sLDS) as natural building blocks for discrete and continuous modeling--which become partially-observable Markov decision processes (POMDPs) and controlled sLDS when augmented with actions. This modular approach supports both passive modeling (generation, forecasting) and active control (planning, decision-making) within the same architecture. We avoid the combinatorial explosion of traditional structure learning by largely fixing the causal architecture and searching over only four depth parameters. We review practical expressiveness through multimodal generative modeling (passive) and planning from pixels (active), with performance competitive to neural approaches while maintaining interpretability. The core outstanding challenge is scalable joint structure-parameter learning; current methods finesse this by cleverly growing structure and parameters incrementally, but are limited in their scalability. If solved, these natural building blocks could provide foundational infrastructure for world modeling, analogous to how standardized layers enabled progress in deep learning.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑