Research map / Vision-language and multimodal / Multimodal LLMs
Reinforcement Learning for Multimodal Diffusion Language Models via Bidimensional Trajectory and Thought Optimization
Bowen Li, Yinjie Wang, Yunzhi Zhang, Junhong Liu, Yingqing Guo, Jiajun Wu, Mengdi Wang, Ling Yang
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add trajectory-level optimization (ablation) | Bi-VRL on LLaDA-V · MMMU | accuracy: 51.6 -> 53.4 | Improved |
| add trajectory-level optimization (ablation) | Bi-VRL on LLaDA-V · SeedBench | accuracy: 78.8 -> 81.1 | Improved |
| add trajectory-level optimization (ablation) | Bi-VRL on MMaDA · MathVerse | accuracy: 26.6 -> 28.3 | Improved |
| add fine-grained multimodal alignment (ablation) | Bi-VRL on LLaDA-V · MMStar | accuracy: 61.6 -> 63.2 | Improved |
| add fine-grained multimodal alignment (ablation) | Bi-VRL on LLaDA-V · MMMU | accuracy: 52.7 -> 53.4 | Improved |
| add fine-grained multimodal alignment (ablation) | Bi-VRL on MMaDA · MMMU | accuracy: 32.6 -> 33.8 | Improved |
| add thought-level optimization (ablation) | Bi-VRL on LLaDA-V · MathVerse | accuracy: 30.9 -> 31.2 | Improved |
| add thought-level optimization (ablation) | Bi-VRL on MMaDA · MathVerse | accuracy: 27.8 -> 28.3 | Improved |
| add thought-level optimization (ablation) | Bi-VRL on LLaDA-V · SeedBench | accuracy: 81.4 -> 81.1 | Hurt |
| add thought-level optimization (ablation) | Bi-VRL on MMaDA · MMBench | accuracy: 68.8 -> 68.1 | Hurt |
| use Bi-VRL instead of Random Masking RL | Bi-VRL on LLaDA-V · MathVerse | accuracy: 28.8 -> 31.2 | Improved |
| use Bi-VRL instead of Coupled-GRPO | Bi-VRL on LLaDA-V · MathVerse | accuracy: 29.2 -> 31.2 | Improved |
Closest papers by meaning
- MMaDA: Multimodal Large Diffusion Language Models (NeurIPS 2025 · similarity 0.92)
- Thinking Diffusion: Penalize and Guide Visual-Grounded Reasoning in Diffusion Multimodal Language Models (CVPR 2026 · similarity 0.91)
- Lavida-R1: Advancing Reasoning for Unified Multimodal Diffusion Language Models (ICML 2026 · similarity 0.90)
- I Think, Therefore I Diffuse: Enabling Multimodal In-Context Reasoning in Diffusion Models (ICML 2025 · similarity 0.90)
- Bifrost-1: Bridging Multimodal LLMs and Diffusion Models with Patch-level CLIP Latents (NeurIPS 2025 · similarity 0.90)
- Revolutionizing Reinforcement Learning Framework for Diffusion Large Language Models (ICLR 2026 · similarity 0.89)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
