Metraipolis › archive › dc_f3fef7d7dc67
UniCAR-RL: Seeing Better before Thinking Deeper in Visual Mathematics — arXiv:2609.13849
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.13849
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-15T22:47:24.065Z
- SHA-256 of the archived text
818b1511eadc7a8c3bb46078c0bd1a12380fe2cebb72becbe3b64684e3a5f6c7- Stable URL of this page
https://metraipolis.com/documento/dc_f3fef7d7dc67
How to cite
UniCAR-RL: Seeing Better before Thinking Deeper in Visual Mathematics — arXiv:2609.13849 — arXiv (en), https://arxiv.org/abs/2609.13849 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_f3fef7d7dc67 (Metraipolis archive dc_f3fef7d7dc67, sha256 818b1511eadc7a8c3bb46078c0bd1a12380fe2cebb72becbe3b64684e3a5f6c7)
BibTeX
@misc{dcf3fef7d7dc67,
title = {UniCAR-RL: Seeing Better before Thinking Deeper in Visual Mathematics — arXiv:2609.13849},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.13849},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_f3fef7d7dc67; sha256 818b1511eadc7a8c3bb46078c0bd1a12380fe2cebb72becbe3b64684e3a5f6c7},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.13849. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Multimodal Large Language Models (MLLMs) often struggle with complex mathematical visual reasoning primarily due to a lack of fine-grained perception, causing initial visual hallucinations to directly trigger cascading reasoning failures. In traditional end-to-end reinforcement learning (RL), sparse rewards fail to decouple perceptual hallucinations from logical missteps, hindering targeted perception optimization. Alternatively, fine-tuning with perception-enhanced CoT data incurs high costs and hallucinations. In this paper, we address these challenges by proposing UniCAR-RL, an annotation-free RL framework. By explicitly decoupling the optimization of perception and reasoning during the training process, it achieves isolation and optimization of both capabilities. Specifically, UniCAR-RL consists of three synergistic branches: 1) a Caption-RL branch that optimizes perception capabilities through verifier-guided reasoning validation; 2) a Reasoning-RL branch that performs logical reasoning based on a gold image description to halt cascading errors; 3) a QA-RL branch that retains native end-to-end alignment to ensure robust question-answering performance. Experiments show that UniCAR-RL substantially improves MLLMs' mathematical and visual reasoning using only raw short-answer data. Furthermore, it demonstrates strong generalization across diverse architectures and scales. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-15 in cs.AI). Categorie arXiv: Artificial Intelligence (cs.AI). Pagina del lavoro: https://arxiv.org/abs/2609.13849 — 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 2026-10-15T22:29:03.152Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-15?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).