aiqbench analysis
UniCAR-RL introduces a novel reinforcement learning framework that decouples visual perception from logical reasoning to mitigate hallucinations in multimodal models tackling mathematical tasks. By separating these capabilities into distinct training branches, the method achieves robust performance without requiring expensive annotated datasets or suffering from cascading errors common in end-to-end approaches. This advancement is significant for improving the reliability of MLLMs in complex visual-mathematical reasoning scenarios.
UniCAR-RL: Seeing Better before Thinking Deeper in Visual Mathematics
arXiv:2609.13849v1 Announce Type: new Abstract: 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.
Abstract: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.
Artificial Intelligence (cs.AI)
arXiv:2609.13849 [cs.AI]
(or arXiv:2609.13849v1 [cs.AI] for this version)
https://doi.org/10.48550/arXiv.2609.13849 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 267 words
Related coverage
Humanity’s Last Invention — Richard Socher of Recursive
Latent Space · Sep 14, 2026
The J-Space Debate, Agent Swarms, and Pacing Frontier AI - Digital Minds Newsletter #4
LessWrong · Sep 18, 2026
Game Design Creativity evals
LessWrong · Oct 12, 2026
Research directions in condensation: varieties of objectivity
AI Alignment Forum · Jul 28, 2026