# Ψ-0.6统一世界模型：自适应思维流实现零样本空间推理与反事实仿真

> 原文链接：https://robotics.hellobike.com/papers/2026-10-23-2610-32177-160c614b
> 频道：papers | 更新日期：2026-10-23 | 标签：auto, arxiv, hf-daily-papers
> 核心摘要：Ψ-0.6提出统一视觉语言世界模型，通过自适应思维流交织图像、文本与几何运动表征。该模型在多视角空间推理上零样本超越主流开源模型，支持反事实仿真，并经任务反馈强化学习优化中间表征选择，提升具身物理推理能力。

# Ψ-0.6统一世界模型：自适应思维流实现零样本空间推理与反事实仿真

::: v-pre
> 本文由 AI 根据公开内容整理，可能有误，请以原文为准：[Ψ-0.6: Adaptive Streams of Thought in a Unified Vision-Language World Model](https://arxiv.org/abs/2610.32177)（HF Daily Papers · arXiv 2610.32177）

## 核心内容

针对AI物理推理能力割裂难题，研究团队发布Ψ-0.6统一视觉语言世界模型。该模型引入“思维流”机制，在共享推理过程中自适应组合感知、语言、几何与仿真能力。实验表明，Ψ-0.6在多视角空间推理任务上零样本性能超越领先开源模型，且通过几何中间件推理获得进一步增益，为具身智能提供自适应物理推理新范式。

技术上，Ψ-0.6通过共享Token接口生成并解析图像、文本及显式几何运动表征，支持如“移动物体是否阻挡路径”的反事实仿真。系统采用基于任务反馈的强化学习，联合优化中间表征的选择、生成与应用，显著提升视觉问答准确率并迁移至未见基准。这种让模型自主选择推理表征的机制，解决了专用模型难以组合适配的问题。

## 关键看点

- **思维流架构**：在共享推理过程中自适应交织图像、文本与几何运动表征，通过统一Token接口实现多模态生成与解析
- **空间推理性能**：在多视角空间推理任务上零样本超越领先开源模型，利用几何中间件推理进一步提升精度
- **反事实仿真能力**：支持生成并解析想象场景以推演物理后果，如判断物体移动是否造成路径遮挡
- **RL优化表征选择**：利用任务反馈强化学习联合优化中间表征的选择与生成，VQA准确率提升并迁移至未见基准

## 工程与产业观察

Ψ-0.6验证了统一世界模型通过自适应选择几何与仿真表征进行物理推理的有效性。其RL优化中间件机制为具身大模型提供了从被动预测向主动规划演进的技术路径，对解决复杂操作任务中的长程推理与泛化瓶颈具有重要工程参考价值。

## 英文摘要 (Abstract)

Humans reason about the physical world by flexibly combining different forms of inference as a problem unfolds. Achieving this flexibility is a major open challenge in AI: capabilities often remain separated across specialized models, limiting how they can be composed and jointly adapted to a task. We introduce Ψ-0.6, a unified vision-language world model that adaptively composes perception, language, geometry, and simulation within a shared reasoning process we call streams of thought. These streams interleave images, text, and explicit representations of geometry and motion, all of which the model can generate and interpret through a shared token interface. Ψ-0.6 surpasses leading open-weight models zero-shot on multi-view spatial reasoning, with further gains from reasoning through geometric intermediates. It also supports counterfactual simulation: generating and interpreting imagined scenes to determine, for example, whether moving an object would obstruct a pathway. Reinforcement learning from task feedback jointly optimizes the selection, generation, and use of intermediate representations, improving visual question-answering accuracy with gains that transfer to unseen benchmarks. Ψ-0.6 offers a path toward adaptive physical reasoning -- not only predicting the physical world but also choosing the representations through which to reason.

Hugging Face Daily Papers 热度：5 upvotes
作者：Khai Loong Aw, Klemen Kotar, Baihan Zhang, Yifan Liu, Lilian Chen, Atlas Kazemian, Rahul Venkatesh, Daniel Yamins
:::

## 原始参考来源 (Sources)

- [HF Daily Papers · arXiv 2610.32177：Ψ-0.6: Adaptive Streams of Thought in a Unified Vision-Language World Model](https://arxiv.org/abs/2610.32177)