Pith. sign in

REVIEW 6 cited by

MMCOMPOSITION: Revisiting the Compositionality of Pre-trained Vision-Language Models

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2410.09733 v1 pith:DA44ELRU submitted 2024-10-13 cs.CV

classification cs.CV
keywords compositionalityvlmsmmcompositionreasoningunderstandingvisualabilityacross
verification ladder T0 review T1 audit T2 compute T3 formal
0 comments
read the original abstract

The advent of large Vision-Language Models (VLMs) has significantly advanced multimodal understanding, enabling more sophisticated and accurate integration of visual and textual information across various tasks, including image and video captioning, visual question answering, and cross-modal retrieval. Despite VLMs' superior capabilities, researchers lack a comprehensive understanding of their compositionality -- the ability to understand and produce novel combinations of known visual and textual components. Prior benchmarks provide only a relatively rough compositionality evaluation from the perspectives of objects, relations, and attributes while neglecting deeper reasoning about object interactions, counting, and complex compositions. However, compositionality is a critical ability that facilitates coherent reasoning and understanding across modalities for VLMs. To address this limitation, we propose MMCOMPOSITION, a novel human-annotated benchmark for comprehensively and accurately evaluating VLMs' compositionality. Our proposed benchmark serves as a complement to these earlier works. With MMCOMPOSITION, we can quantify and explore the compositionality of the mainstream VLMs. Surprisingly, we find GPT-4o's compositionality inferior to the best open-source model, and we analyze the underlying reasons. Our experimental analysis reveals the limitations of VLMs in fine-grained compositional perception and reasoning, and points to areas for improvement in VLM design and training. Resources available at: https://hanghuacs.github.io/MMComposition/

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 6 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. MemoBench: Benchmarking World Modeling in Dynamically Changing Environments

    cs.CV 2026-06 unverdicted novelty 7.0 of 10

    None of ten tested video-generation models reliably remembers objects after occlusion in dynamic scenes; static-camera videos inflate consistency scores.

  2. Point-RFT: Improving Multimodal Reasoning with Visually Grounded Reinforcement Finetuning

    cs.CV 2025-05 reject novelty 7.0 of 10

    Point-RFT uses point-grounded chain-of-thought with GRPO reinforcement to improve chart reasoning, reporting 90.04% on ChartQA, though internal tables and the OOD setup weaken the claim.

  3. VideoArgus: Agentic Rubric-Grounded Unified Evaluation for Video Generation and Editing

    cs.CV 2026-08 conditional novelty 6.0 of 10

    An output-blind, rubric-grounded evaluator and benchmark that ranks generated videos closer to human preferences than five existing benchmark-specific evaluators.

  4. Evaluating Compositional Generalisation in VLMs and Diffusion Models

    cs.CV 2025-08 conditional novelty 6.0 of 10

    On the extended Concept Binding Benchmark, all tested VLMs, including the generative Diffusion Classifier, fail to distinguish left/right relations in generalised zero-shot settings, suggesting models rely on object r...

  5. Impact of Pretraining Word Co-occurrence on Compositional Generalization in Multimodal Models

    cs.CV 2025-07 conditional novelty 6.0 of 10

    The accuracy of CLIP and CLIP-based visual question answering models is strongly correlated with how often the concept pair in an image appears together in pretraining captions.

  6. MMIG-Bench: Towards Comprehensive and Explainable Evaluation of Multi-Modal Image Generation Models

    cs.CV 2025-05 conditional novelty 6.0 of 10

    MMIG-Bench is a unified benchmark of 4,850 prompts and 1,750 reference images with a three-level evaluation suite, including the VQA-based Aspect Matching Score that correlates with human ratings.

Pith tools