Guardado en:
Detalles Bibliográficos
Autores principales: Bai, Jinbin, Lei, Yu, Wu, Hecong, Zhu, Yuchen, Li, Shufan, Xin, Yi, Li, Xiangtai, Tao, Molei, Grover, Aditya, Yang, Ming-Hsuan
Formato: Preprint
Publicado: 2025
Materias:
Acceso en línea:https://arxiv.org/abs/2510.20668
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866908607648366592
author Bai, Jinbin
Lei, Yu
Wu, Hecong
Zhu, Yuchen
Li, Shufan
Xin, Yi
Li, Xiangtai
Tao, Molei
Grover, Aditya
Yang, Ming-Hsuan
author_facet Bai, Jinbin
Lei, Yu
Wu, Hecong
Zhu, Yuchen
Li, Shufan
Xin, Yi
Li, Xiangtai
Tao, Molei
Grover, Aditya
Yang, Ming-Hsuan
contents This is not a typical survey of world models; it is a guide for those who want to build worlds. We do not aim to catalog every paper that has ever mentioned a ``world model". Instead, we follow one clear road: from early masked models that unified representation learning across modalities, to unified architectures that share a single paradigm, then to interactive generative models that close the action-perception loop, and finally to memory-augmented systems that sustain consistent worlds over time. We bypass loosely related branches to focus on the core: the generative heart, the interactive loop, and the memory system. We show that this is the most promising path towards true world models.
format Preprint
id arxiv_https___arxiv_org_abs_2510_20668
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle From Masks to Worlds: A Hitchhiker's Guide to World Models
Bai, Jinbin
Lei, Yu
Wu, Hecong
Zhu, Yuchen
Li, Shufan
Xin, Yi
Li, Xiangtai
Tao, Molei
Grover, Aditya
Yang, Ming-Hsuan
Machine Learning
This is not a typical survey of world models; it is a guide for those who want to build worlds. We do not aim to catalog every paper that has ever mentioned a ``world model". Instead, we follow one clear road: from early masked models that unified representation learning across modalities, to unified architectures that share a single paradigm, then to interactive generative models that close the action-perception loop, and finally to memory-augmented systems that sustain consistent worlds over time. We bypass loosely related branches to focus on the core: the generative heart, the interactive loop, and the memory system. We show that this is the most promising path towards true world models.
title From Masks to Worlds: A Hitchhiker's Guide to World Models
topic Machine Learning
url https://arxiv.org/abs/2510.20668