Towards a Data-Centric Understanding of Spatial Reasoning in Multimodal Large Language Models

Title: Towards a Data-Centric Understanding of Spatial Reasoning in Multimodal Large Language Models

标题：迈向对多模态大语言模型中空间推理的数据中心化理解

Computer Science > Computer Vision and Pattern Recognition arXiv:2610.19108 (cs) [Submitted on 13 Oct 2026] Title:Towards a Data-Centric Understanding of Spatial Reasoning in Multimodal Large Language Models Authors:Mahdieh Mallahnezhad, Masoud Jafaripour, Mohamadali Tofigh, Sheila Schoepp, Osmar R. Zaiane

计算机科学 > 计算机视觉与模式识别 arXiv:2610.19108 (cs) [提交于 2026 年 10 月 13 日] 标题：迈向对多模态大语言模型中空间推理的数据中心化理解 作者：Mahdieh Mallahnezhad, Masoud Jafaripour, Mohamadali Tofigh, Sheila Schoepp, Osmar R. Zaiane

Abstract: Multimodal Large Language Models (MLLMs) have been widely applied in embodied agents and robotics for vision tasks. However, MLLMs suffer from limitations in spatial reasoning and perception. Post-training techniques have shown promise for improving these capabilities; yet, there are limited data-driven surveys of such approaches.

摘要：多模态大语言模型（MLLMs）已被广泛应用于具身智能体和机器人的视觉任务中。然而，MLLMs 在空间推理和感知方面存在局限性。训练后技术在提升这些能力方面展现出了前景；然而，针对此类方法的数据驱动型综述仍然有限。

In this survey, we consider approaches to fine-tuning pretrained MLLMs for spatial reasoning tasks, ranging from supervised to reinforcement learning, with an emphasis on the datasets used. We investigate themes of dataset construction, data modality and annotation, and evaluation of the resulting models. Based on these themes, we categorize research into three groups: dataset design, dataset construction and evaluation.

在本综述中，我们探讨了针对空间推理任务微调预训练 MLLMs 的方法，范围涵盖从监督学习到强化学习，并重点关注所使用的数据集。我们调查了数据集构建、数据模态与标注，以及所得模型的评估等主题。基于这些主题，我们将研究分为三类：数据集设计、数据集构建和评估。

We find four main methods for constructing fixed spatial annotations: rules and templates, LLM prompting, procedural generation, and human-in-the-loop. We also observe substantial variation in evaluation approaches, including different scoring methods, tolerance metrics, and benchmark types. Finally, we identify opportunities for future research in unified dynamic annotation, efficient spatial reasoning, standardization of evaluation, expanded spatial range, and human-centered spatial interactions.

我们发现了构建固定空间标注的四种主要方法：规则与模板、LLM 提示、程序化生成以及人在回路。我们还观察到评估方法存在显著差异，包括不同的评分方法、容差指标和基准类型。最后，我们确定了未来研究的机会，包括统一动态标注、高效空间推理、评估标准化、扩展空间范围以及以人为本的空间交互。