Papers with spatial reasoning
CAPruner: Conceptual-Adjacent Scene Graph Pruner for Enhancing 3D Spatial Reasoning of Large Language Models (2026.acl-long)
Copied to clipboard
| Challenge: | Existing pruning methods rely on spatial proximity and remove relevant relations, thereby undermining reliable spatial reasoning. |
| Approach: | They propose a scene graph pruning model that integrates fuzzy semantic relevance with spatial proximity to estimate the importance of relations. |
| Outcome: | Experiments show that CAPruner outperforms proximity-based pruning with negligible cost savings. |
Action Inference for Destination Prediction in Vision-and-Language Navigation (2024.acl-srw)
Copied to clipboard
| Challenge: | Existing work on vision-and-language navigation focuses on spatial reasoning and semantic grounding of visual information, but there is still scope for improvement. |
| Approach: | They propose a VLN task of destination prediction for picking up a pedestrian that requires action inference from a crowd-sourced dataset. |
| Outcome: | The proposed model can reason about the effect of the next action and the next on the destination to a certain extent. |
MapAgent: A Hierarchical Agent for Geospatial Reasoning with Dynamic Map Tool Integration (2026.findings-eacl)
Copied to clipboard
| Challenge: | Existing frameworks for large language models are tailored to domains such as mathematics, coding, or web automation. |
| Approach: | They propose a hierarchical multi-agent plug-and-play framework with customized toolsets and agentic scaffolds for map-integrated geospatial reasoning. |
| Outcome: | The proposed framework decouples planning from execution and reduces cognitive load on users. |
FlowVQA: Mapping Multimodal Logic in Visual Question Answering with Flowcharts (2024.findings-acl)
Copied to clipboard
Shubhankar Singh, Purvi Chaurasia, Yerram Varun, Pranshu Pandya, Vatsal Gupta, Vivek Gupta, Dan Roth
| Challenge: | Existing benchmarks for visual question answering lack in visual grounding and complexity, particularly in evaluating spatial reasoning skills. |
| Approach: | They propose to use flowcharts as visual contexts to assess the capabilities of visual question-answering multimodal language models in reasoning. |
| Outcome: | The proposed benchmarks evaluate models' ability to follow visual information without pre-existing knowledge on a suite of open-source and proprietary multimodal language models using various strategies, followed by an analysis of directional bias. |
Using Perspectival Words Is Harder Than Vocabulary Words for Humans —and Even More So for Multimodal Language Models (2026.acl-long)
Copied to clipboard
| Challenge: | Existing evaluations of multimodal language models focus on vocabulary words with relatively stable, context-independent meanings in conversation, such as object names, colors, and verbs. |
| Approach: | They compare human and multimodal language models in their use of three word types: vocabulary, possessives, and demonstratives. |
| Outcome: | The models approach human-level performance on using vocabulary, but exhibit clear deficits with possessives and even greater difficulties with demonstratives. |
GAMED.AI: A Hierarchical Multi-Agent Framework for Automated Educational Game Generation (2026.acl-demo)
Copied to clipboard
| Challenge: | GameDAI is a hierarchical multi-agent framework that transforms instructor-provided questions into fully playable, pedagogically grounded educational games validated through formal mechanic contracts. |
| Approach: | They propose a hierarchical multi-agent framework that transforms instructor-provided questions into fully playable, pedagogically grounded educational games validated through formal mechanic contracts. |
| Outcome: | The proposed framework achieves 90% validation pass rate, 98.3% schema compliance, and 73% token reduction over ReAct agents at 0.46 per game. |
TopViewRS: Vision-Language Models as Top-View Spatial Reasoners (2024.emnlp-main)
Copied to clipboard
| Challenge: | Top-view perspective is a typical way in which humans read and reason over different types of maps, but spatial reasoning capabilities of modern VLMs in this setup remain unattested and underexplored. |
| Approach: | They introduce a top-view spatial reasoning dataset and use it to evaluate VLMs across 4 perception and reasoning tasks with different levels of complexity. |
| Outcome: | The proposed model can understand and reason over spatial relations from the top view and can be controlled at different granularities of spatial reasoning. |
Evaluation of Multilingual Ability to Use Spatial Deictic Expressions in Vision-Language Models (2026.acl-srw)
Copied to clipboard
| Challenge: | Existing studies have focused on the ability of vision-language models to utilize spatial deictic expressions, which depend on the situation of utterance. |
| Approach: | They develop a benchmark to evaluate the multilingual ability of VLMs to use spatial deictic expressions in four languages. |
| Outcome: | The proposed models use demonstratives in a different manner from humans, particularly in selecting demonstrative based on distance from the object. |
Neuro-symbolic Training for Reasoning over Spatial Language (2025.findings-naacl)
Copied to clipboard
| Challenge: | Spatial reasoning is essential for everyday human tasks and is crucial for robots to interact with their environment in a human-like manner. |
| Approach: | They propose to train language models to adhere to spatial reasoning rules as constraints . this allows them to capture the necessary level of abstraction for spatial reasoning . |
| Outcome: | The proposed technique improves language models in multi-hop spatial reasoning over text . it achieves higher accuracy than other competitive Spatial Question-answering benchmarks . |
The Power of Bullet Lists: A Simple Yet Effective Prompting Approach to Enhancing Spatial Reasoning in Large Language Models (2025.findings-naacl)
Copied to clipboard
| Challenge: | Large language models (LLMs) are currently dominating the field of natural language processing, but spatial reasoning ability is lacking in LLMs. |
| Approach: | They propose a prompting technique that integrates bullet lists, coordinates, and visualizations into the reasoning process and integrates them into planning tasks. |
| Outcome: | The proposed technique boosts LLMs' spatial reasoning abilities compared to previous prompting techniques. |
Sparkle: Mastering Basic Spatial Capabilities in Vision Language Models Elicits Generalization to Spatial Reasoning (2025.findings-emnlp)
Copied to clipboard
Yihong Tang, Ao Qu, Zhaokai Wang, Dingyi Zhuang, Zhaofeng Wu, Wei Ma, Shenhao Wang, Yunhan Zheng, Zhan Zhao, Jinhua Zhao
| Challenge: | Currently, vision-language models excel in many downstream tasks but struggle with spatial reasoning, which is crucial for navigation and interaction with physical environments. |
| Approach: | They propose a framework that generates synthetic data to provide targeted supervision for VLMs across these basic spatial capabilities. |
| Outcome: | The proposed framework disentangles 2D spatial reasoning into three core components: direction comprehension, distance estimation, and localization. |
Disentangling Extraction and Reasoning in Multi-hop Spatial Reasoning (2023.findings-emnlp)
Copied to clipboard
| Challenge: | Recent studies highlight the struggles even large language models encounter when it comes to performing spatial reasoning over text. |
| Approach: | They propose to disentangle spatial reasoning over text and compare them to state-of-the-art models with no explicit design for these parts. |
| Outcome: | The proposed models show that they can perform spatial reasoning over text and can generalize within real data domains. |
DART: Leveraging Multi-Agent Disagreement for Tool Recruitment in Multimodal Reasoning (2026.eacl-long)
Copied to clipboard
Nithin Sivakumaran, Justin Chen, David Wan, Yue Zhang, Jaehong Yoon, Elias Stengel-Eskin, Mohit Bansal
| Challenge: | a key strength of human intelligence is the ability to debate and discuss reasoning with others. |
| Approach: | They propose a multi-agent framework that uses disagreements between visual agents to identify useful visual tools that can resolve inter-agency disagreement. |
| Outcome: | The proposed framework beats the strongest baseline on A-OKVQA and MMMU, respectively. |
SpaRC and SpaRP: Spatial Reasoning Characterization and Path Generation for Understanding Spatial Reasoning Capability of Large Language Models (2024.acl-long)
Copied to clipboard
| Challenge: | Existing large language models (LLMs) do not perform well on the datasets. |
| Approach: | They propose to use a Spatial Reasoning Characterization framework and a spatial reasoning path framework to study spatial reasoning. |
| Outcome: | The proposed framework and datasets outperform state-of-the-art models in spatial reasoning. |
LLM-Driven Multi-Perspective Location Completion for Next Location Prediction (2026.findings-acl)
Copied to clipboard
| Challenge: | Existing methods assume that check-in data is complete, overlooking the subjective nature of user behavior, leading to inaccurate capture of user preferences. |
| Approach: | They propose a framework that uses spatial coordinates to augment location completion by transforming geographic coordinates into text. |
| Outcome: | The proposed framework outperforms state-of-the-art methods on three real-world datasets. |
TURTLEAI: Benchmarking Multimodal Models for Visual Programming in Turtle Graphics (2026.findings-acl)
Copied to clipboard
| Challenge: | Vision-language models have been explored for visual programming, but performance is unclear . most prior work focuses on visual programming for productivity . |
| Approach: | They propose a visual programming benchmark that uses visual programming to evaluate VLMs. |
| Outcome: | The proposed model improves on GPT-5, GPT-4o, and Qwen2-VL-72B on real-world tasks by 20% . the proposed model is based on 823 visual programming tasks in the Turtle Graphics domain . |
Bridging the Dynamic Perception Gap: Training-Free Draft Chain-of-Thought for Dynamic Multimodal Spatial Reasoning (2025.findings-emnlp)
Copied to clipboard
| Challenge: | Existing methods for dynamic spatial reasoning are limited to text or static visual domains . |
| Approach: | They propose a framework that augments textual reasoning chains with dynamic visual drafts . |
| Outcome: | The proposed framework outperforms existing methods in dynamic spatial reasoning tasks. |
SPARTQA: A Textual Question Answering Benchmark for Spatial Reasoning (2021.naacl-main)
Copied to clipboard
| Challenge: | Existing studies have focused on the spatial reasoning capabilities of modern language models (LMs) however, there has been limited research into the spatial thinking capabilities of LMs. |
| Approach: | They propose a question-answering (QA) benchmark for spatial reasoning on natural language text which contains more realistic spatial phenomena not covered by prior work. |
| Outcome: | The proposed method significantly improves LMs' ability on spatial understanding, which in turn helps solve two external datasets, bAbI, and boolQ. |
SpaRE: Enhancing Spatial Reasoning in Vision-Language Models with Synthetic Data (2025.acl-long)
Copied to clipboard
| Challenge: | Vision-language models struggle with spatial reasoning, a skill that humans excel at. |
| Approach: | They propose to use a spatial-reasoning Enhanced (SpaRE) VLM to improve spatial reasoning in visual question answering and robotics. |
| Outcome: | The proposed model achieves a 49% performance gain on the What's Up benchmark while maintaining strong results on general tasks. |
Can MLLMs Find Their Way in a City? Exploring Emergent Navigation from Web-Scale Knowledge (2026.eacl-long)
Copied to clipboard
| Challenge: | Existing evaluation benchmarks for multimodal large language models (MLLMs) are language-centric or heavily reliant on simulated environments, rarely probing the nuanced, knowledge-intensive reasoning essential for practical, real-world scenarios. |
| Approach: | They propose a task of Sparsely Grounded Visual Navigation to evaluate MLLM-driven agents in city navigation in four diverse global cities. |
| Outcome: | The proposed benchmark encompassing four diverse global cities evaluates agents' decision-making abilities in city navigation. |
DepWiGNN: A Depth-wise Graph Neural Network for Multi-hop Spatial Reasoning in Text (2023.findings-emnlp)
Copied to clipboard
| Challenge: | Existing approaches for spatial reasoning in text overlook the gap between natural language and symbolic structures. |
| Approach: | They propose a novel depth-wise Graph Neural Network to aggregate spatial information over the depth dimension instead of the breadth dimension of the graph. |
| Outcome: | The proposed model outperforms existing methods on two multi-hop spatial reasoning datasets. |
LOViS: Learning Orientation and Visual Signals for Vision and Language Navigation (2022.coling-1)
Copied to clipboard
| Challenge: | Existing Transformer-based VLN agents entangle orientation and vision information, which limits the learning of each information source. |
| Approach: | They propose to design a navigation agent with explicit Orientation and Vision modules . they use a set of pre-training tasks to feed the modules into the model . |
| Outcome: | The proposed model improves on R2R and R4R datasets and achieves state-of-the-art results. |
SPaRC: A Spatial Pathfinding Reasoning Challenge (2025.emnlp-main)
Copied to clipboard
| Challenge: | Existing reasoning datasets saturate and fail to test abstract, multi-step problems, especially pathfinding and complex rule constraint satisfaction. |
| Approach: | They propose to use a spatial few-shot grid to evaluate spatial and rule-based reasoning with 1,000 2D grid puzzles. |
| Outcome: | The proposed model can be used to evaluate spatial reasoning and improve its accuracy. |
CityEQA: A Hierarchical LLM Agent on Embodied Question Answering Benchmark in City Space (2025.emnlp-main)
Copied to clipboard
Yong Zhao, Kai Xu, Zhengqiu Zhu, Yue Hu, Zhiheng Zheng, Yingfeng Chen, Yatai Ji, Chen Gao, Yong Li, Jincai Huang
| Challenge: | Embodied Question Answering (EQA) tasks are primarily focused on indoor environments, leaving the complexities of urban settings unexplored. |
| Approach: | They propose a task where an embodied agent answers open-vocabulary questions in dynamic city spaces. |
| Outcome: | The proposed agent achieves 60.7% of human-level answering accuracy compared to baselines . the proposed agent outperforms existing agents in open-ended city spaces . |
AlanaVLM: A Multimodal Embodied AI Foundation Model for Egocentric Video Understanding (2024.findings-emnlp)
Copied to clipboard
Alessandro Suglia, Claudio Greco, Katie Baker, Jose Part, Ioannis Papaioannou, Arash Eshghi, Ioannis Konstas, Oliver Lemon
| Challenge: | Current Vision-Language Models (VLMs) focus on third-person view videos, neglecting the richness of egocentric perceptual experience. |
| Approach: | They propose to use the Egocentric Video Understanding Dataset (EVUD) to train VLMs on video captioning and question answering tasks specific to egocentric videos. |
| Outcome: | The proposed model outperforms open-source models including strong Socratic models using GPT-4 as a planner by 3.6% and outperformed Claude 3 and Gemini Pro Vision 1.0. |
Emma-X: An Embodied Multimodal Action Model with Grounded Chain of Thought and Look-ahead Spatial Reasoning (2025.acl-long)
Copied to clipboard
| Challenge: | Visual-Language-Action models lack the ability to generate actionable policies tailored to specific robotic embodiments. |
| Approach: | They propose an embodied multimodal action model with Grounded Chain of Thought and Look-ahead Spatial Reasoning that enhances spatial reasoning and task planning. |
| Outcome: | The proposed model improves on existing baselines in tasks requiring spatial reasoning and grounding reasoning. |
LLMArena: Assessing Capabilities of Large Language Models in Dynamic Multi-Agent Environments (2024.acl-long)
Copied to clipboard
| Challenge: | Existing benchmarks for evaluating large language models use static datasets, leading to data leakage or overlooking the complexities of multi-agent interactions. |
| Approach: | They propose a framework that evaluates the diverse capabilities of LLM agents in multi-agent dynamic environments. |
| Outcome: | The proposed framework assesses the diverse capabilities of LLM agents in multi-agent dynamic environments. |
Vision-and-Language Navigation with Analogical Textual Descriptions in LLMs (2025.emnlp-main)
Copied to clipboard
| Challenge: | Existing zero-shot LLM-based Vision-and-Language Navigation agents either encode images as textual scene descriptions, potentially oversimplifying visual details, or process raw image inputs, which can fail to capture abstract semantics required for high-level reasoning. |
| Approach: | They propose to integrate large language models into embodied AI models by incorporating textual descriptions that facilitate analogical reasoning across images from multiple perspectives. |
| Outcome: | The proposed approach improves the agent’s contextual understanding on the R2R dataset, showing that it can make better decisions based on the LLMs. |
CADMate: Generating CAD Assembly Plan with Geometric Chain-of-Thought and Spatial Physical Rewards (2026.acl-long)
Copied to clipboard
| Challenge: | Computer-aided design (CAD) is crucial in prototyping complex 3D objects . designers manually define assembly sequences for individual CAD parts . |
| Approach: | They propose a framework for computer-aided design that predicts actions for CAD parts . they use a reference design image and disassembled parts to generate 6-DoF transformations . |
| Outcome: | The proposed framework outperforms existing MLLMs in the design of CAD assemblies. |
Knowing More, Acting Better: Hierarchical Representation for Embodied Decision-Making (2025.findings-emnlp)
Copied to clipboard
| Challenge: | Modern embodied AI uses multimodal large language models as policy models, predicting actions from final-layer hidden states. |
| Approach: | They propose a hierarchical action probing method that aggregates representations from all layers, mirroring the brain's multi-level organization. |
| Outcome: | Experiments show that hierarchical probing improves on last-layer embodied models and achieves a 46.6% success rate and a 62.5% gain in spatial reasoning tasks. |
Ascending the Infinite Ladder: Benchmarking Spatial Deformation Reasoning in Vision-Language Models (2026.acl-long)
Copied to clipboard
Jiahuan Zhang, Shunwen Bai, Tianheng Wang, KaiWen Guo, Zijia Song, Hanqing WU, Guozheng Rao, Kai Han, Kaicheng Yu
| Challenge: | Existing benchmarks explore aspects of threedimensional spatial reasoning and visual-language reasoning in dynamic environments, but they are unable to perform well on 3D spatial deformation reasoning. |
| Approach: | They propose to use a ladder competition format to assess the model's spatial deformation reasoning abilities to determine its performance. |
| Outcome: | The proposed framework assesses the performance of Vision-Language Models in spatial deformation reasoning tasks. |
An Empirical Analysis on Spatial Reasoning Capabilities of Large Multimodal Models (2024.emnlp-main)
Copied to clipboard
| Challenge: | Large Multimodal Models (LMMs) have shown impressive generalization ability on vision and language tasks, but their spatial understanding is under-explored. |
| Approach: | They construct a VQA dataset to analyze LMMs' spatial reasoning capabilities. |
| Outcome: | The proposed model is stronger at basic object detection than complex spatial reasoning. |
NUMINA: A Natural Understanding Benchmark for Multi-dimensional Intelligence and Numerical Reasoning Abilities (2025.findings-emnlp)
Copied to clipboard
Changyu Zeng, Yifan Wang, Zimu Wang, Wei Wang, Zhengni Yang, Muyi Bao, Jimin Xiao, Anh Nguyen, Yutao Yue
| Challenge: | Existing 3D benchmarks lack fine-grained numerical reasoning task annotations, limiting MLLMs’ ability to perform precise spatial measurements and complex numerical reasoning. |
| Approach: | They propose a 3D-based benchmark to enhance indoor perceptual understanding by using multi-scale annotations and question-answer pairs. |
| Outcome: | The proposed benchmark improves indoor perceptual understanding by incorporating multi-scale annotations and question-answer pairs. |
VLN-MME: Diagnosing MLLMs as Language-guided Visual Navigation Agents (2026.acl-long)
Copied to clipboard
| Challenge: | Multimodal Large Language Models have demonstrated remarkable capabilities across vision-language tasks, but their performance as embodied agents needs further exploration. |
| Approach: | They propose a framework to evaluate multimodal large language models as zero-shot agents . they find that enhancing prevalent agents with Chain-of-Thought reasoning and self-reflection leads to an unexpected performance decrease. |
| Outcome: | The proposed framework enables comparisons and component-level ablations across diverse MLLM architectures, agent designs, and navigation tasks. |
iVISPAR — An Interactive Visual-Spatial Reasoning Benchmark for VLMs (2025.emnlp-main)
Copied to clipboard
| Challenge: | Vision-Language Models (VLMs) struggle with spatial reasoning and visual alignment, despite their performance on 2D tasks. |
| Approach: | They propose a multimodal benchmark to evaluate VLMs' spatial reasoning capabilities based on the sliding tile puzzle . |
| Outcome: | The proposed model performs better on 2D tasks compared to 3D or text-based settings, but struggles with complex spatial configurations and consistently falls short of human performance. |
One Cognitive Loop Is Enough: SODA unlocks Pure-Text Spatial Reasoning in Large Language Models (2026.acl-long)
Copied to clipboard
Shunwen Bai, Jiahuan Zhang, Haoran Huang, Yurun Wang, Jiale Liu, Yanxi Wu, Ningzhe Yu, Yudong Gao, Mingjun Cheng
| Challenge: | Existing large language models (LLMs) lack visual input, leading to errors in basic numerical comparisons. |
| Approach: | They propose a spatial OODA framework that integrates the OODAC cognitive loop into multiple control tasks and integrates it into LLMs. |
| Outcome: | The proposed model significantly improves the spatial reasoning capabilities of large language models across multiple scenarios including SPOD-Bench, SPACE and applications. |
CoV: Chain-of-View Prompting for Spatial Reasoning (2026.findings-acl)
Copied to clipboard
Haoyu Zhao, Akide Liu, Zeyu Zhang, Weijie Wang, Feng Chen, Ruihan Zhu, Gholamreza Haffari, Bohan Zhuang
| Challenge: | Embodied question answering requires collecting context that is distributed across multiple viewpoints . most recent vision–language models (VLMs) are constrained to a fixed and finite set of input views . |
| Approach: | They propose a training-free, test-time reasoning framework that transforms a VLM into an active viewpoint reasoner through a coarse-to-fine exploration process. |
| Outcome: | The proposed framework improves LLM-Match performance by 11.98% on four mainstream VLMs. |
Embodied-Reasoner: Synergizing Visual Search, Reasoning, and Action for Embodied Interactive Tasks (2026.acl-long)
Copied to clipboard
Wenqi Zhang, Mengna Wang, Gangao Liu, Huixin Xu, Yiwei Jiang, Yongliang Shen, Guiyang Hou, Zhe Zheng, Hang Zhang, Xin Li, Jiajun Liu, Weiming Lu, Peng Li, Yueting Zhuang
| Challenge: | Recent advances in reasoning models have demonstrated remarkable capabilities on mathematical and coding tasks, but their effectiveness in embodied domains remains largely unexplored. |
| Approach: | They propose a reasoning model for interactive embodied tasks that synthesizes 9.3k coherent Observation-Thought-Action trajectories containing 64k ego-centric images and 90k diverse reasoning processes. |
| Outcome: | The proposed model outperforms existing visual reasoning models by +9%, 24%, and +13% on long-horizon tasks. |