Metraipolis › archive › dc_009c62bd989d
GroundSight at GroundLM 2026 Shared Tasks: GoldenViewVQA — arXiv:2610.11402
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2610.11402
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-09T06:30:31.316Z
- SHA-256 of the archived text
f86733fda3dcd82418411ba380ceb17841f4c99fd4133dc14b8d148da782ed5c- Stable URL of this page
https://metraipolis.com/documento/dc_009c62bd989d
How to cite
GroundSight at GroundLM 2026 Shared Tasks: GoldenViewVQA — arXiv:2610.11402 — arXiv (en), https://arxiv.org/abs/2610.11402 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_009c62bd989d (Metraipolis archive dc_009c62bd989d, sha256 f86733fda3dcd82418411ba380ceb17841f4c99fd4133dc14b8d148da782ed5c)
BibTeX
@misc{dc009c62bd989d,
title = {GroundSight at GroundLM 2026 Shared Tasks: GoldenViewVQA — arXiv:2610.11402},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2610.11402},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_009c62bd989d; sha256 f86733fda3dcd82418411ba380ceb17841f4c99fd4133dc14b8d148da782ed5c},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.11402. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
GoldenViewVQA requires models to jointly answer driving-scene questions and identify the camera view containing the supporting visual evidence, making precise evidence localization as important as answer correctness. We present \textbf{CoVeR-VQA}, a training-free multi-stage verification and correction framework for grounded multi-view VQA. Starting from GPT-5.6 zero-shot predictions, CoVeR-VQA progressively applies view-specific verification with Gemini-3.6-Flash, prior-guided joint verification with Claude-Opus-5, and cross-split group-level verification that exploits semantically filtered question groups from shared multi-view scenes and validation-derived prior knowledge. On the official GoldenViewVQA test set, the four-stage CoVeR-VQA pipeline achieves 84.75\% Joint Accuracy, improving the GPT-5.6 zero-shot baseline by 13.56 percentage points, while reaching 94.92\% Answer Accuracy and 86.44\% View Accuracy. The final submitted run achieves 88.14\% Joint Accuracy after two additional evaluator-informed post-hoc corrections. Our analysis shows that supporting-view localization remains the primary source of residual errors, highlighting the importance of explicit evidence verification for reliable multi-view multimodal reasoning.
Tema degli esploratori: statistica e scienza dei dati (elenco arXiv cs.CV, nuovi lavori del giorno).
Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV).
Pagina del lavoro: https://arxiv.org/abs/2610.11402 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-09T05:55:07.388Z dalla pagina https://arxiv.org/list/cs.CV/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).