{"ID":2899646,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2507.00748","arxiv_id":"2507.00748","title":"Improving the Reasoning of Multi-Image Grounding in MLLMs via Reinforcement Learning","abstract":"Multimodal Large Language Models (MLLMs) perform well in single-image visual grounding but struggle with real-world tasks that demand cross-image reasoning and multi-modal instructions. To address this, we adopt a reinforcement learning (RL) based post-training strategy for MLLMs in multi-image grounding tasks. We first synthesize high-quality chain-of-thought (CoT) data for cold-start initialization, followed by supervised fine-tuning (SFT) using low-rank adaptation (LoRA). Subsequently, we apply rejection sampling with the merged SFT model to curate reliable RL data and use rule-based RL to guide the model toward optimal reasoning paths. Extensive experiments demonstrate the effectiveness of our approach, achieving +9.04% on MIG-Bench and +4.41% on average across seven out-of-domain benchmarks.","short_abstract":"Multimodal Large Language Models (MLLMs) perform well in single-image visual grounding but struggle with real-world tasks that demand cross-image reasoning and multi-modal instructions. To address this, we adopt a reinforcement learning (RL) based post-training strategy for MLLMs in multi-image grounding tasks. We firs...","url_abs":"https://arxiv.org/abs/2507.00748","url_pdf":"https://arxiv.org/pdf/2507.00748v3","authors":"[\"Bob Zhang\",\"Haoran Li\",\"Tao Zhang\",\"Jianan Li\",\"Cilin Yan\",\"Xikai Liu\",\"Jiayin Cai\",\"Yanbin Hao\"]","published":"2025-07-01T13:48:57Z","proceeding":"cs.CV","tasks":"[\"cs.CV\"]","methods":"[\"Reinforcement Learning\",\"Large Language Model\",\"Language Model\",\"LoRA\"]","has_code":false}