cs.CVNov 18, 2026

GTA: Generative Tower of Agents for Scalable Holonic Scene Generation

Authors: Zeng Tao, Yunuo Chen, Xiang Feng, Yumeng He, Peihao Li, Nanxiang Jiang, Yaowei Guo, Fanlian Zeng, 

Organizations: UCLA · Envora · UC Berkeley · University of Utah

Abstract

Scalable world modeling requires coordinating scene-wide goals with fine-grained objects and dynamic characters under limited time and context budgets. We introduce GTA (Generative Tower of Agents), which represents every unit at every level of an adaptive scene hierarchy as a holon, both a self-contained construction task and a composable scene unit. Parallel agents recursively plan the hierarchy top-down, generate assets, and assemble and refine them bottom-up. Uniform interfaces carry scene-wide requirements down the hierarchy, guiding local construction so that each holon fits into the larger one. GTA accelerates scalable world modeling through high-throughput parallel execution while preserving well-structured layouts and fine-grained asset details. GTA then populates the assembled environments with dynamic human characters whose appearance and motion are guided by the scene, yielding realized 4D worlds. Project page: https://zt-zeng-tao.github.io/GTA/

Figures & tables

Explore similar work

CardsList
  1. SceneAgent: A Multi-Agent Framework for Portable Game Scene Generation

    Nov 13, 2026Xingyi Li, Jiahao Yang, Yiqin Zhu +3Game Generation3D Scene Generation

  2. WorldClaw: Agentic 3D Open-World Generation at Scale

    Aug 5, 2026Chunchao Guo, Jinpeng Li, Yang Li +1Open-Ended Generation3D Asset Generation

  3. Gamma-World: Generative Multi-Agent World Modeling Beyond Two Players

    May 27, 2026Fangfu Liu, Kai He, Tianchang Shen +7Interactive World ModelsVideo World Models