Research map / Vision-language and multimodal / Multimodal LLMs
MoAI: Mixture of All Intelligence for Large Language and Vision Models
Byung-Kwan Lee, Beomchan Park, Chae Won Kim, Yong Man Ro
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 panoptic segmentation and open-world detection outputs (ablation) | MoAI · MME Existence | score: 187 -> 200 | Improved |
| add scene graph generation outputs (ablation) | MoAI · MME Position | score: 145 -> 165 | Improved |
| add OCR model outputs (ablation) | MoAI · MM-Bench OCR | accuracy: 80.1 -> 93.5 | Improved |
| combine both training steps instead of the first only (ablation) | MoAI · MME-P | score: 1542 -> 1714 | Improved |
| use top-3 experts instead of top-1 (ablation) | MoAI-Mixer · MME-P | score: 1588 -> 1714 | Improved |
| train gating network instead of uniform weights (ablation) | MoAI-Mixer · MME-P | score: 1617 -> 1714 | Improved |
| add bounding boxes and segmentation map together (ablation) | MoAI · MM-Bench Localization | accuracy: 55.2 -> 71.1 | Improved |
| add auxiliary CV information to other LLVMs | LLaVA1.5-7B · MM-Bench | accuracy: 64.3 -> 79.1 | Improved |
| add auxiliary CV information to other LLVMs | Qwen-VL-Chat-7B · TextVQA | accuracy: 63.8 -> 69.2 | Improved |
| increase number of detection categories from 0 to 1847 (ablation) | MoAI · MM-Bench | accuracy: 65.9 -> 79.3 | Improved |
| MoAI versus larger CogVLM-17B | MoAI-7B · MM-Vet | score: 43.7 vs 54.5 | Hurt |
Later papers that compared with this method
1 later accepted paper reports MoAI-7B as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.
- R2-T2: Re-Routing in Test-Time for Multimodal Mixture-of-Experts (ICML 2025 · better on 9 of 9 shared results)
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 2 on most shared results.
- CogVLM-17B — CogVLM: Visual Expert for Pretrained Language Models (NeurIPS 2024) · better on 0 of 1 shared result
- ShareGPT4V-7B — ShareGPT4V: Improving Large Multi-Modal Models with Better Captions (ECCV 2024) · better on 1 of 1 shared result
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- MoCHA: Advanced Vision-Language Reasoning with MoE Connector and Hierarchical Group Attention (AAAI 2026 · similarity 0.91)
- MoVA: Adapting Mixture of Vision Experts to Multimodal Context (NeurIPS 2024 · similarity 0.89)
- Jack of All Tasks Master of Many: Designing General-Purpose Coarse-to-Fine Vision-Language Model (CVPR 2024 · similarity 0.88)
- Benchmarking Large Vision-Language Models on Fine-Grained Image Tasks: A Comprehensive Evaluation (ICLR 2026 · similarity 0.88)
- MMLongBench: Benchmarking Long-Context Vision-Language Models Effectively and Thoroughly (NeurIPS 2025 · similarity 0.88)
- Are We on the Right Way for Evaluating Large Vision-Language Models? (NeurIPS 2024 · similarity 0.88)
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.
