用大模型先验引导流模型，让智能体在未知环境里更准想象物体位置。
Distilling LLM Prior to Flow Model for Generalizable Agent's Imagination in Object Goal Navigation
- 用大模型生成场景语义图，转为二维高斯场注入流模型作为空间先验
- 在MP3D和Gibson上达到当前最优，跨数据集迁移表现强
- 适合做通用导航的智能体研发，尤其关注泛化能力的研究者
物体目标导航（ObjectNav）任务要求智能体在未见环境中定位指定物体，需对未观测区域进行场景想象。现有方法依赖确定性判别模型构建语义地图，忽视了室内布局的固有不确定性，限制了在未见环境中的泛化能力。本文提出GOAL，一种基于生成流的框架，通过将大语言模型（LLM）增强的完整场景语义图与观测区域连接，建模室内环境的语义分布。训练时，从大语言模型推断的空间先验被编码为二维高斯场，并注入目标地图，将丰富的上下文知识蒸馏到流模型中，实现更通用的场景补全。大量实验表明，GOAL在MP3D和Gibson数据集上达到当前最优性能，并在向HM3D数据集的迁移设置中展现强大泛化能力。代码与预训练模型已开源：https://github.com/Badi-Li/GOAL。
原文摘要 · Abstract (English)
The Object Goal Navigation (ObjectNav) task challenges agents to locate a specified object in an unseen environment by imagining unobserved regions of the scene. Prior approaches rely on deterministic and discriminative models to complete semantic maps, overlooking the inherent uncertainty in indoor layouts and limiting their ability to generalize to unseen environments. In this work, we propose GOAL, a generative flow-based framework that models the semantic distribution of indoor environments by bridging observed regions with LLM-enriched full-scene semantic maps. During training, spatial priors inferred from large language models (LLMs) are encoded as two-dimensional Gaussian fields and injected into target maps, distilling rich contextual knowledge into the flow model and enabling more generalizable completions. Extensive experiments demonstrate that GOAL achieves state-of-the-art performance on MP3D and Gibson, and shows strong generalization in transfer settings to HM3D. Codes and pretrained models are available at https://github.com/Badi-Li/GOAL.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。