Improving the Reasoning of Multi-Image Grounding in MLLMs via Reinforcement Learning
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866913023163105280 |
|---|---|
| author | Zhang, Bob Li, Haoran Zhang, Tao Li, Jianan Yan, Cilin Liu, Xikai Cai, Jiayin Hao, Yanbin |
| author_facet | Zhang, Bob Li, Haoran Zhang, Tao Li, Jianan Yan, Cilin Liu, Xikai Cai, Jiayin Hao, Yanbin |
| contents | 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. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2507_00748 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Improving the Reasoning of Multi-Image Grounding in MLLMs via Reinforcement Learning Zhang, Bob Li, Haoran Zhang, Tao Li, Jianan Yan, Cilin Liu, Xikai Cai, Jiayin Hao, Yanbin Computer Vision and Pattern Recognition 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. |
| title | Improving the Reasoning of Multi-Image Grounding in MLLMs via Reinforcement Learning |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2507.00748 |