Can you SPLICE it together? A Human Curated Benchmark for Probing Visual Reasoning in VLMs
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866911183697608704 |
|---|---|
| author | Ballout, Mohamad Wilfred, Okajevo Yaghoubi, Seyedalireza Abdelmoneim, Nohayr Muhammad Mayer, Julius Bruni, Elia |
| author_facet | Ballout, Mohamad Wilfred, Okajevo Yaghoubi, Seyedalireza Abdelmoneim, Nohayr Muhammad Mayer, Julius Bruni, Elia |
| contents | In this work, we introduce SPLICE, a human-curated benchmark derived from the COIN instructional video dataset, designed to probe event-based reasoning across multiple dimensions: temporal, causal, spatial, contextual, and general knowledge. SPLICE includes 3,381 human-filtered videos spanning 12 categories and 180 sub-categories, such as sports, engineering, and housework. These videos are segmented into a total of 11,423 event clips. We evaluate both human participants and state-of-the-art vision-language models (VLMs) on the task of rearranging these clips into coherent event sequences to assess visual reasoning capabilities. Results reveal a significant gap: VLMs struggle to match human performance. While human-annotated textual descriptions improve model accuracy, they do not affect human performance, suggesting that models rely more on language priors than on visual understanding. Even with annotations, VLMs fall short of human-level reasoning, underscoring persistent challenges in visual reasoning. A deeper analysis across sub-categories shows that VLMs perform relatively better on videos where temporal and causal reasoning are dominant, compared to those where contextual and spatial reasoning are dominant. They also perform better on everyday tasks than on specialized ones. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2509_24640 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Can you SPLICE it together? A Human Curated Benchmark for Probing Visual Reasoning in VLMs Ballout, Mohamad Wilfred, Okajevo Yaghoubi, Seyedalireza Abdelmoneim, Nohayr Muhammad Mayer, Julius Bruni, Elia Computer Vision and Pattern Recognition Artificial Intelligence In this work, we introduce SPLICE, a human-curated benchmark derived from the COIN instructional video dataset, designed to probe event-based reasoning across multiple dimensions: temporal, causal, spatial, contextual, and general knowledge. SPLICE includes 3,381 human-filtered videos spanning 12 categories and 180 sub-categories, such as sports, engineering, and housework. These videos are segmented into a total of 11,423 event clips. We evaluate both human participants and state-of-the-art vision-language models (VLMs) on the task of rearranging these clips into coherent event sequences to assess visual reasoning capabilities. Results reveal a significant gap: VLMs struggle to match human performance. While human-annotated textual descriptions improve model accuracy, they do not affect human performance, suggesting that models rely more on language priors than on visual understanding. Even with annotations, VLMs fall short of human-level reasoning, underscoring persistent challenges in visual reasoning. A deeper analysis across sub-categories shows that VLMs perform relatively better on videos where temporal and causal reasoning are dominant, compared to those where contextual and spatial reasoning are dominant. They also perform better on everyday tasks than on specialized ones. |
| title | Can you SPLICE it together? A Human Curated Benchmark for Probing Visual Reasoning in VLMs |
| topic | Computer Vision and Pattern Recognition Artificial Intelligence |
| url | https://arxiv.org/abs/2509.24640 |