ZimaBlue: Evolving Generalizable World Action Models through Scalable Video Pre-training
- 类型：arxiv
- 标识：2609.00188
- 链接：https://arxiv.org/abs/2609.00188
- 主分类：multimodal
- 形态：method
- TLDR：Robotic manipulation faces a fundamental scaling challenge: robust generalization demands broad physical experience, yet action-labeled robot trajectories are expensive to collect and inherently limited in diversity. Egocentric videos offer a far more scalable source of embodied experience, capturing object interactions, contact dynamics, tool use, and long-horizon behaviors across diverse environments. The central challenge is how to convert this abundant but action-free experience into effective robot control. We introduce ZimaBlue, a scalable framework for learning generalizable World Actio
- 副分类：engineering
- 待LLM分类：否
- 来源文件：
- /inbox/tom/_candidates/2026-09-02-agent-rag-longcontext-candidates.json