追求空间推理的最小充分性
Pursuing Minimal Sufficiency in Spatial Reasoning
AI总结：
本文提出 MSSR 框架，通过构建最小充分集提升视觉-语言模型的空间推理能力，实现充分性与最小性的平衡，取得最佳性能并生成可解释的推理路径。
摘要：
空间推理，即能够将语言 grounding 在 3D 理解的能力，仍然是视觉-语言模型 (VLMs) 中持续存在的挑战。我们识别出两个根本性的瓶颈：源于 2D 中心预训练的 3D 理解能力不足，以及由冗余 3D 信息引起的推理失败。为了解决这些问题，我们首先在回答给定问题之前构建一个最小充分集 (MSS) 的信息：从专家模型中选取紧凑的 3D 感知结果。我们引入 MSSR (最小充分空间推理器)，一种双代理框架，实现了这一原则。感知代理通过多功能感知工具箱程序性地查询 3D 场景，提取足够的信息，包括一个新颖的 SOG (Situated Orientation Grounding) 模块，能够稳健地提取语言 grounding 的方向。推理代理随后迭代地细化这些信息以追求最小性，修剪冗余细节并请求缺失的信息，在闭环中直到 MSS 被精心编纂。广泛的实验表明，我们的方法通过显式追求充分性和最小性，显著提高了准确性，并在两个具有挑战性的基准上实现了最先进的性能。此外，我们的框架生成了可解释的推理路径，为未来模型提供了有前景的高质量训练数据来源。源代码可在 https://github.com/gyj155/mssr 上获得。
英文摘要：
Spatial reasoning, the ability to ground language in 3D understanding, remains a persistent challenge for Vision-Language Models (VLMs). We identify two fundamental bottlenecks: inadequate 3D understanding capabilities stemming from 2D-centric pre-training, and reasoning failures induced by redundant 3D information. To address these, we first construct a Minimal Sufficient Set (MSS) of information before answering a given question: a compact selection of 3D perception results from \textit{expert models}. We introduce MSSR (Minimal Sufficient Spatial Reasoner), a dual-agent framework that implements this principle. A Perception Agent programmatically queries 3D scenes using a versatile perception toolbox to extract sufficient information, including a novel SOG (Situated Orientation Grounding) module that robustly extracts language-grounded directions. A Reasoning Agent then iteratively refines this information to pursue minimality, pruning redundant details and requesting missing ones in a closed loop until the MSS is curated. Extensive experiments demonstrate that our method, by explicitly pursuing both sufficiency and minimality, significantly improves accuracy and achieves state-of-the-art performance across two challenging benchmarks. Furthermore, our framework produces interpretable reasoning paths, offering a promising source of high-quality training data for future models. Source code is available at https://github.com/gyj155/mssr.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
