研究人员正在开发新方法来提高视觉语言模型（VLMs）的空间推理能力，这些模型目前在3D理解方面存在困难。几篇论文提出将几何先验和结构化推理注入这些模型。技术包括使用正交视图、几何感知空间先验（GASP）、分层任务分解和点云数据来提高在空间基准测试上的性能。 AI
影响 这些进展旨在为AI模型提供更强大的3D空间理解能力，这对于机器人和具身智能等应用至关重要。
排序理由 多篇学术论文提出改进AI模型能力的新方法和基准测试。
在 Hugging Face Daily Papers 阅读 →
AI 生成摘要 · Google Gemini · 来自 35 个来源。 我们如何撰写摘要 →
arXiv:2603.07751v2 Announce Type: replace-cross Abstract: Current Large Language Models have achieved Olympiad-level logic, yet Vision-Language Models paradoxically falter on elementary spatial tasks like block counting. This capability mismatch reveals a critical ``spatial intel…
arXiv:2605.30231v1 Announce Type: cross Abstract: Vision-Language Models (VLMs) often struggle with robust 3D spatial reasoning. Prevailing methods that rely on fine-tuning with 3D visual question-answering (VQA) datasets may overfit dataset-specific biases, while integrating spe…
Vision-Language Models (VLMs) often struggle with robust 3D spatial reasoning. Prevailing methods that rely on fine-tuning with 3D visual question-answering (VQA) datasets may overfit dataset-specific biases, while integrating specialized 3D visual encoders is often inflexible an…
Vision-Language Models (VLMs) often struggle with robust 3D spatial reasoning. Prevailing methods that rely on fine-tuning with 3D visual question-answering (VQA) datasets may overfit dataset-specific biases, while integrating specialized 3D visual encoders is often inflexible an…
arXiv:2601.03048v2 Announce Type: replace-cross Abstract: Vision Transformers (ViTs) excel in semantic recognition but exhibit systematic failures in spatial reasoning tasks such as mental rotation. While often attributed to data scale, this work argues that the limitation arises…
arXiv:2605.28144v1 Announce Type: new Abstract: LLMs have shown remarkable proficiency in general language understanding and reasoning. However, they consistently underperform in spatial reasoning that severely limits their application, particularly in embodied intelligence. Insp…
arXiv:2605.28277v1 Announce Type: new Abstract: Whether large language models (LLMs) construct internal spatial world models from pure-text descriptions remains contested, and whether such capabilities transfer across languages has not been systematically studied. We introduce Me…
arXiv:2605.28490v1 Announce Type: cross Abstract: 3D object grounding localizes referred objects in a 3D scene from natural language. Unified instance-centric 3D-LLMs aim to solve grounding together with dialog, QA, and captioning, yet many rely on a single pointer-style groundin…
arXiv:2504.04540v2 Announce Type: replace-cross Abstract: 3D Large Language Models (LLMs) leveraging spatial information in point clouds for 3D spatial reasoning attract great attention. Despite some promising results, the advantages of point clouds over other modalities remain u…
Training Vision-Language Models with geometric priors improves 3D spatial reasoning through deep supervision with contrastive loss and depth consistency, achieving better performance than standard fine-tuning approaches.
arXiv:2509.21552v2 Announce Type: replace-cross Abstract: Graphical User Interface (GUI) grounding is commonly framed as a coordinate prediction task -- given a natural language instruction, generate on-screen coordinates for actions such as clicks and keystrokes. However, recent…
arXiv:2507.07644v4 Announce Type: replace Abstract: We introduce FloorplanQA, a diagnostic benchmark for evaluating spatial reasoning in large language models (LLMs). FloorplanQA is grounded in structured representations of indoor scenes, such as (e.g., kitchens, living rooms, be…
arXiv:2505.23764v3 Announce Type: replace-cross Abstract: Spatial intelligence is essential for multimodal large language models (MLLMs) operating in the complex physical world. Existing benchmarks, however, probe only single-image relations and thus fail to assess the multi-imag…
Training intervention called View Dropout combined with panoramic visual thinking enables more effective cross-view spatial reasoning in unified multimodal models.
In this work, we study Cooperative Spatial Intelligence, the ability of decentralized embodied agents to coordinate effectively under dynamic environmental constraints across city-scale outdoor domains. We introduce Sentinel Challenge, a benchmark where multiple decentralized emb…
arXiv:2505.17015v2 Announce Type: replace-cross Abstract: Multi-modal large language models (MLLMs) have rapidly advanced in visual tasks, yet their spatial understanding remains limited to single images, leaving them ill-suited for physical-world applications that require multi-…
arXiv:2605.20837v1 Announce Type: cross Abstract: Architectural spatial intelligence, the ability to recognize and infer architectural space, is fundamental to tasks such as robot navigation, embodied interaction, and 3D scene understanding and generation. Although extensive rese…
arXiv:2605.21625v1 Announce Type: cross Abstract: The emergence of Large Vision-Language Models (LVLMs) has significantly advanced video understanding capabilities. However, existing benchmarks focus predominantly on coarse-grained tasks such as action segmentation, classificatio…
arXiv:2605.22536v1 Announce Type: cross Abstract: Multimodal Large Language Models (MLLMs) have made rapid progress in spatial intelligence, yet existing spatial reasoning benchmarks largely assume pristine visual inputs and overlook the degradations that commonly occur in real-w…
Multimodal Large Language Models (MLLMs) have made rapid progress in spatial intelligence, yet existing spatial reasoning benchmarks largely assume pristine visual inputs and overlook the degradations that commonly occur in real-world deployment, such as motion blur, low light, a…
SpaceDG dataset and benchmark evaluate multimodal language models' spatial reasoning robustness under visual degradations, revealing significant performance gaps and demonstrating improved robustness through targeted training.
Architectural spatial intelligence, the ability to recognize and infer architectural space, is fundamental to tasks such as robot navigation, embodied interaction, and 3D scene understanding and generation. Although extensive research has evaluated the basic spatial skills of Vis…
Spatial intelligence requires multimodal large language models (MLLMs) to move beyond single-view perception and reason consistently about objects, visibility, geometry, and interactions across multiple viewpoints. However, progress in cross-view reasoning remains limited by thre…
arXiv:2605.29074v1 Announce Type: new Abstract: Are current Vision Language Models (VLMs) ready to comprehend and reason about complex embodied interactions in 3D environments? We introduce Embodied3DBench, a robot-centric benchmark targeting low-level spatial intelligence in emb…
arXiv:2605.28132v1 Announce Type: new Abstract: Spatial intelligence requires visual representations that capture both semantic objects and geometric structure in the physical world. To support this, two major pre-training schemes are now widely used as foundation backbones: Visi…
3D object grounding localizes referred objects in a 3D scene from natural language. Unified instance-centric 3D-LLMs aim to solve grounding together with dialog, QA, and captioning, yet many rely on a single pointer-style grounding decision that compresses a relational instructio…
Spatial intelligence requires visual representations that capture both semantic objects and geometric structure in the physical world. To support this, two major pre-training schemes are now widely used as foundation backbones: Vision-Language Models (VLMs), which use language su…
arXiv:2605.27310v1 Announce Type: new Abstract: Cross-view spatial reasoning remains a weak spot for vision-language models (VLMs): they often reason in language and lose the fine-grained geometry needed for the task. Thinking with images aims to address this by generating an int…
arXiv:2605.26239v1 Announce Type: new Abstract: In this work, we study Cooperative Spatial Intelligence, the ability of decentralized embodied agents to coordinate effectively under dynamic environmental constraints across city-scale outdoor domains. We introduce Sentinel Challen…
arXiv:2510.09606v2 Announce Type: replace Abstract: With the current surge in spatial reasoning explorations, researchers have made significant progress in understanding indoor scenes, but still struggle with diverse applications such as robotics and autonomous driving. This pape…
Cross-view spatial reasoning remains a weak spot for vision-language models (VLMs): they often reason in language and lose the fine-grained geometry needed for the task. Thinking with images aims to address this by generating an intermediate thinking image, but recent work shows …
arXiv:2604.03318v2 Announce Type: replace Abstract: Multimodal large language models (MLLMs) are increasingly being applied to spatial cognition tasks, where they are expected to understand and interact with complex environments. Most existing works improve spatial reasoning by i…
arXiv:2605.25524v1 Announce Type: new Abstract: Reliable spatial reasoning remains a core bottleneck for vision-language models (VLMs). Existing mainstream training paradigms for spatial reasoning largely rely on outcome alignment or process imitation, lacking explicit constraint…
arXiv:2605.25334v1 Announce Type: new Abstract: Spatial understanding of the physical world from 2D visual inputs hinges on two complementary forms of geometric knowledge: holistic 3D structural perception and fine-grained metric scale estimation. Existing multimodal large langua…
Vision-Language Models (VLMs) have made striking progress, yet their spatial reasoning remains fragile: models that answer an original input correctly can still fail under paired transformations with predictable answer mappings, revealing a gap between instance-level correctness …