What’s “up” with vision-language models? Investigating their struggle with spatial reasoning (2023.emnlp-main)
Copied to clipboard
| Challenge: | Recent work has re-surfaced a concern that has long plagued vision-language models: poor performance on simple tasks like attribute attachment, counting, etc. |
| Approach: | They evaluate 18 vision-language models and find they perform poorly on VQAv2 . they find that popular vision-linguistic pretraining corpora lack reliable data for learning spatial relationships . |
| Outcome: | The new models are compared with existing datasets on what'sup and visual-language models . they achieve 56% accuracy on the new benchmarks compared to 99% for humans . |
Similar Papers
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. |
Visual Spatial Reasoning (2023.tacl-1)
Copied to clipboard
| Challenge: | Existing benchmarks for testing vision-language models (VLMs) are not ideal as they conflate multiple sources of error and do not allow controlled analysis on specific linguistic or cognitive properties. |
| Approach: | They present a dataset containing more than 10k natural text-image pairs with 66 types of spatial relations in English (e.g., under, in front of, facing). |
| Outcome: | The proposed model fails to capture relational information in a visual question answering task and referring expression comprehension tasks. |
SPHERE: Unveiling Spatial Blind Spots in Vision-Language Models Through Hierarchical Evaluation (2025.acl-long)
Copied to clipboard
Wenyu Zhang, Wei En Ng, Lixin Ma, Yuwen Wang, Junqi Zhao, Allison Koenecke, Boyang Li, Wanglu Wanglu
| Challenge: | Current vision-language models lack multi-dimensional spatial reasoning capabilities for human-like understanding and applications. |
| Approach: | They propose a hierarchical evaluation framework that probes models across increasing levels of complexity and integrates spatial, visual, and logical understanding. |
| Outcome: | The proposed framework probes models across increasing levels of complexity, from basic skills to multi-skill integration and high-level reasoning that combines spatial, visual, and logical understanding. |
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. |
Measuring Progress in Fine-grained Vision-and-Language Understanding (2023.acl-long)
Copied to clipboard
| Challenge: | X-VLM models lack "fine-grained" understanding of relationships, verbs and numbers in images . pretraining on large-scale image–text data from the Web has facilitated rapid progress on many vision-and-language tasks . |
| Approach: | They investigate models that outperform other baselines on fine-grained data . they highlight importance of novel losses and rich data sources for learning fine-grain skills . |
| Outcome: | The proposed model outperforms baseline models on four fine-grained benchmarks . the model outpersforms other baseline models and even degrades performance . |
Response Wide Shut? Surprising Observations in Basic Vision Language Model Capabilities (2025.acl-long)
Copied to clipboard
| Challenge: | Vision-language Models have been shown to be highly capable but lacking basic visual understanding skills. |
| Approach: | They propose to examine the limitations of vision-language models on visual tasks by constructing a series of tests that probe which components of design may be lacking. |
| Outcome: | The proposed tests compare VLMs to other models on visual encoders, intermediate vision-language projection and LLM-decoder outputs. |
What’s Missing in Vision-Language Models? Probing Their Struggles with Causal Order Reasoning (2026.eacl-long)
Copied to clipboard
| Challenge: | Existing benchmarks often include a mixture of reasoning questions, making it difficult to truly assess VLMs’ causal reasoning abilities. |
| Approach: | They propose two new benchmarks specifically designed to isolate and rigorously evaluate VLMs’ causal reasoning abilities. |
| Outcome: | The proposed benchmarks show that vision-language models perform poorly on causal reasoning tasks, often only marginally surpassing random guessing. |
Can VLMs Actually See and Read? A Survey on Modality Collapse in Vision-Language Models (2025.findings-acl)
Copied to clipboard
| Challenge: | Vision-language models integrate textual and visual information, enabling them to process visual inputs and generate predictions. |
| Approach: | They review work on modality collapse analysis to provide insights into the reason for this unintended behavior and review probing studies for fine-grained vision-language understanding. |
| Outcome: | The proposed models can achieve competitive performance in vision-language tasks despite relying heavily on textual information and ignoring visual information. |
NegVQA: Can Vision Language Models Understand Negation? (2025.findings-acl)
Copied to clipboard
| Challenge: | NegVQA is a visual question answering (VQA) benchmark consisting of 7,379 two-choice questions covering diverse negation scenarios and image-question distributions. |
| Approach: | They propose a visual question answering benchmark consisting of 7,379 two-choice questions covering diverse negation scenarios and image-question distributions. |
| Outcome: | The proposed model fails to correctly interpret negation, leading to critical errors in interactive AI systems. |
On Efficient Language and Vision Assistants for Visually-Situated Natural Language Understanding: What Matters in Reading and Reasoning (2024.emnlp-main)
Copied to clipboard
| Challenge: | Recent advances in language and vision assistants have showcased impressive capabilities but suffer from a lack of transparency, limiting broader research and reproducibility. |
| Approach: | They propose to redefine the design of vision-language models by identifying key components and creating efficient models with constrained inference costs. |
| Outcome: | The proposed models achieve significant improvements in inference throughput while maintaining high performance. |