Research map / Vision-language and multimodal / Multimodal LLMs
From Training-Free to Adaptive: Empirical Insights into MLLMs’ Understanding of Detection Information
Qirui Jiao, Daoyuan Chen, Yilun Huang, Yaliang Li, Ying Shen
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 |
|---|---|---|---|
| infuse textual detection info training-free | LLaVA-1.5-7B · 10 MLLM benchmarks | average improvement: +2.30% | Improved |
| infuse textual detection info training-free | LLaVA-1.5-13B · 10 MLLM benchmarks | average improvement: -3.41% | Hurt |
| retrain with detection info (RBI) | LLaVA-1.5-7B · 10 MLLM benchmarks | average improvement: +1.91% | Improved |
| retrain with detection info (RBI) | LLaVA-1.5-13B · GQA | accuracy: 81.0 -> 78.0 | Hurt |
| fine-tune one extra epoch with detection info (FTBI) | LLaVA-1.5-7B · 10 MLLM benchmarks | average improvement: +3.99% | Improved |
| fine-tune one extra epoch with detection info (FTBI) | LLaVA-1.5-13B · 10 MLLM benchmarks | average improvement: +3.30% | Improved |
| fine-tune instead of retrain (without detection info at inference) (ablation) | LLaVA-1.5-7B · GQA | accuracy: 74.8 -> 78.4 | Improved |
| add detection info at inference to RBI (ablation) | RBI-7B · GQA | accuracy: 74.8 -> 76.6 | Improved |
| add detection info during fine-tuning (ablation) | FTBI-7B · MM-Vet | score: 29.8 -> 35.2 | Improved |
| add detection info during fine-tuning (ablation) | FTBI-13B · MM-Vet | score: 36.1 -> 38.9 | Improved |
| infuse OCR information only (ablation) | FTBI-7B · TextVQA | accuracy: 57.1 -> 60.3 | Improved |
| infuse object detection information only (ablation) | FTBI-7B · POPE | accuracy: 86.0 -> 89.0 | Improved |
Closest papers by meaning
- Divide, Conquer and Combine: A Training-Free Framework for High-Resolution Image Perception in Multimodal Large Language Models (AAAI 2025 · similarity 0.90)
- Benchmarking Large Vision-Language Models on Fine-Grained Image Tasks: A Comprehensive Evaluation (ICLR 2026 · similarity 0.89)
- MLLMs Know Where to Look: Training-free Perception of Small Visual Details with Multimodal LLMs (ICLR 2025 · similarity 0.89)
- ROD-MLLM: Towards More Reliable Object Detection in Multimodal Large Language Models (CVPR 2025 · similarity 0.89)
- Machine Vision Therapy: Multimodal Large Language Models Can Enhance Visual Robustness via Denoising In-Context Learning (ICML 2024 · similarity 0.89)
- Are We on the Right Way for Evaluating Large Vision-Language Models? (NeurIPS 2024 · 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.
