A blank-image ablation test reveals that high probe accuracy on VLM spatial reasoning frequently reflects priors or inverted signs rather than image grounding, with horizontal grounded, vertical prior, and depth inverted.
hub
What’s" up" with vision-language models? investigating their strug- gle with spatial reasoning.arXiv preprint arXiv:2310.19785
11 Pith papers cite this work. Polarity classification is still indexing.
11
Pith papers citing it
hub tools
representative citing papers
PhotoFlow is a closed-loop agent framework that searches for camera parameters in 3D scenes according to language intent and outperforms one-shot, reflection, and random baselines on the new VPhotoBench of 47 scenes and 141 missions.
VisAnalog is a new controlled benchmark showing VLMs substantially underperform humans on visual concept transfer under one- to four-step deterministic transformations, with relation inference as the main failure mode.
Proposes Spatial Narrative Score (SNS) evaluation for VLMs' camera motion understanding and introduces CaMo model achieving consistent performance on SNS and direct QA.
GeoWorld-VLM aligns VLM image features with intermediate representations from camera-conditioned world models via fine-tuning only the encoder and projector, yielding ~4% gains on What'sUp and VSR spatial benchmarks across two VLM backbones.
SpatialStack improves 3D spatial reasoning in vision-language models by stacking and synchronizing multi-level geometric features with the language backbone.
Introduces group matching score for better evaluation of compositional reasoning and Test-Time Matching (TTM) algorithm for unsupervised self-improvement in multimodal models, achieving SOTA gains including surpassing GPT-4.1 and estimated human performance.
Embodied-R1 uses a pointing-centric representation and reinforced fine-tuning on a 200K dataset to achieve state-of-the-art results on embodied benchmarks plus 56.2% success in SIMPLEREnv and 87.5% on real XArm tasks without task-specific training.
VILASR integrates visual drawing operations with reasoning in LVLMs via cold-start synthetic training, reflective rejection sampling, and reinforcement learning, yielding an 18.4% average gain on spatial reasoning benchmarks.
Empirical study shows bidirectional but sensitive relationship between compositionality and long-caption understanding in VLMs, promoted by high-quality grounded data and affected by architectural choices like frozen positional embeddings.
AutoSpatial improves VLM spatial reasoning for social navigation by combining minimal manual supervision with auto-labeled VQA pairs and hierarchical training, showing gains up to 20.5% in action prediction over baselines.
citing papers explorer
-
Decodable Is Not Grounded: A Vision-Ablation Arbiter for VLM Spatial Reasoning
A blank-image ablation test reveals that high probe accuracy on VLM spatial reasoning frequently reflects priors or inverted signs rather than image grounding, with horizontal grounded, vertical prior, and depth inverted.
-
PhotoFlow: Agentic 3D Virtual Photography Missions
PhotoFlow is a closed-loop agent framework that searches for camera parameters in 3D scenes according to language intent and outperforms one-shot, reflection, and random baselines on the new VPhotoBench of 47 scenes and 141 missions.
-
VisAnalog: A Diagnostic Suite for Visual Concept Transfer on Natural Images
VisAnalog is a new controlled benchmark showing VLMs substantially underperform humans on visual concept transfer under one- to four-step deterministic transformations, with relation inference as the main failure mode.
-
CaMo: Camera Motion Grounded Evaluation and Training for Vision-Language Models
Proposes Spatial Narrative Score (SNS) evaluation for VLMs' camera motion understanding and introduces CaMo model achieving consistent performance on SNS and direct QA.
-
GeoWorld-VLM: Geometry from World Models for Vision-Language Models
GeoWorld-VLM aligns VLM image features with intermediate representations from camera-conditioned world models via fine-tuning only the encoder and projector, yielding ~4% gains on What'sUp and VSR spatial benchmarks across two VLM backbones.
-
SpatialStack: Layered Geometry-Language Fusion for 3D VLM Spatial Reasoning
SpatialStack improves 3D spatial reasoning in vision-language models by stacking and synchronizing multi-level geometric features with the language backbone.
-
Test-Time Matching: Unlocking Compositional Reasoning in Multimodal Models
Introduces group matching score for better evaluation of compositional reasoning and Test-Time Matching (TTM) algorithm for unsupervised self-improvement in multimodal models, achieving SOTA gains including surpassing GPT-4.1 and estimated human performance.
-
Embodied-R1: Reinforced Embodied Reasoning for General Robotic Manipulation
Embodied-R1 uses a pointing-centric representation and reinforced fine-tuning on a 200K dataset to achieve state-of-the-art results on embodied benchmarks plus 56.2% success in SIMPLEREnv and 87.5% on real XArm tasks without task-specific training.
-
Reinforcing Spatial Reasoning in Vision-Language Models with Interwoven Thinking and Visual Drawing
VILASR integrates visual drawing operations with reasoning in LVLMs via cold-start synthetic training, reflective rejection sampling, and reinforcement learning, yielding an 18.4% average gain on spatial reasoning benchmarks.
-
Long Story Short: Disentangling Compositionality and Long-Caption Understanding in Contrastive VLMs
Empirical study shows bidirectional but sensitive relationship between compositionality and long-caption understanding in VLMs, promoted by high-quality grounded data and affected by architectural choices like frozen positional embeddings.
-
AutoSpatial: Visual-Language Reasoning for Social Robot Navigation through Efficient Spatial Reasoning Learning
AutoSpatial improves VLM spatial reasoning for social navigation by combining minimal manual supervision with auto-labeled VQA pairs and hierarchical training, showing gains up to 20.5% in action prediction over baselines.