# 面向 World Action Models 的事件对齐视觉动作推理

> 原文链接：https://robotics.hellobike.com/papers/2026-10-07-2610-09427
> 频道：papers | 更新日期：2026-10-07 | 标签：arxiv, cs.RO
> 核心摘要：提出按交互事件组织视觉预测的框架，用于WAM动作生成，并在DOMINO和RoboTwin 2.0上验证。

# 面向 World Action Models 的事件对齐视觉动作推理

::: v-pre
> 原文链接：[Event-Aligned Visual Action Reasoning for World Action Models](https://arxiv.org/abs/2610.09427)（arXiv 2610.09427 · 发布日期：2026-10-07）

**Event-Aligned Visual Action Reasoning for World Action Models**

提出按交互事件组织视觉预测的框架，用于WAM动作生成，并在DOMINO和RoboTwin 2.0上验证。

## 核心要点

- 现有WAM按预设时间间隔组织视觉想象，未显式区分关键交互与连接过渡的作用。
- 方法围绕交互事件进行视觉-动作预测监督，使想象展开更关注影响动作生成的关键状态变化。
- 引入执行有效性头识别预测动作序列的有效部分，在分块推理中避免冗余动作。
- 实验在DOMINO成功率较基线提升10.26个百分点，并在RoboTwin 2.0表现具竞争力。

**验证范围与局限**：验证范围为DOMINO、RoboTwin 2.0，以及从DOMINO Level 1到Levels 2和3的无目标级适配迁移。

## 英文摘要 (Abstract)

World-Action Models (WAMs) utilize future visual prediction as an intermediate reasoning process to guide action generation. However, existing WAMs typically structure visual imagination according to predefined temporal intervals, without explicitly accounting for the different roles of task-critical interactions and connecting transitions. We argue that effective visual foresight should align directly with task-relevant interactions and their corresponding reasoning demands. To this end, we introduce an event-aligned visual action reasoning framework that organizes visual-action prediction around interaction events. Through event-aligned visual-action supervision, WAM learns to generate event-aligned visual context in each imagined rollout, placing greater emphasis on critical state changes that inform action generation. This shapes the visual reasoning granularity according to the underlying interaction dynamics, with detailed reasoning around task-critical events and coarser progression through connecting transitions. Furthermore, we introduce an execution validity head that identifies the valid portion of each predicted action sequence, avoiding redundant actions during chunked inference. Experiments demonstrate a 10.26 percentage point improvement in DOMINO success rate over baseline and competitive performance on RoboTwin 2.0. It also transfers from DOMINO Level 1 to Levels 2 and 3 without target-level adaptation.

作者：Xiaomeng Yang, Yushu Wu, Yi Gao, Yuhao Lei, Xuan Zhang, Pu Zhao, Yanzhi Wang
:::

## 原始参考来源 (Sources)

- [arXiv 2610.09427：Event-Aligned Visual Action Reasoning for World Action Models](https://arxiv.org/abs/2610.09427)