Research map / Detection and segmentation / Open-set and few-shot detection
MAU-GPT: Enhancing Multi-type Industrial Anomaly Understanding via Anomaly-aware and Generalist Experts Adaptation
Zhuonan Wang, Zhenxuan Fan, Siwen Tan, Yu Zhong, Yuqian Yuan, Haoyuan Li, Hao Jiang, Wenqiao Zhang, Feifei Shao, Hongwei Wang, Jun Xiao
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 |
|---|---|---|---|
| use AMoE-LoRA instead of plain LoRA (ablation) | MAU-GPT 4B · MAU-Set Open-Ended QA | BLEU-4: 0.4529 -> 0.4773 | Improved |
| use AMoE-LoRA instead of LoRAMoE (ablation) | MAU-GPT 4B · MAU-Set Open-Ended QA | ROUGE-1: 0.6991 -> 0.7026 | Improved |
| use AMoE-LoRA instead of plain LoRA (ablation) | MAU-GPT 4B · MAU-Set Open-Ended QA | ROUGE-L: 0.6449 -> 0.6531 | Improved |
| set 16 generalist experts with rank 4 (product fixed at 64) (ablation) | MAU-GPT 4B · MAU-Set Open-Ended QA | ROUGE/BLEU | Improved |
| domain-specific fine-tuning on MAU-Set vs general-purpose VLMs | MAU-GPT 4B · MAU-Set Discriminative QA (Avg.) | accuracy: 73.54 -> 82.12 | Improved |
| add more in-context shots to AnomalyGPT | AnomalyGPT · MAU-Set Discriminative QA (Avg.) | accuracy: 62.53 -> 66.03 | Improved |
| add more in-context shots to AnomalyGPT | AnomalyGPT · MAU-Set Open-Ended QA (Avg., GPT-4o judged) | accuracy: 21.61 -> 22.30 | No clear effect |
| scale LLaVA-1.5 from 7B to 13B | LLaVA-1.5 · MAU-Set Discriminative QA (Avg.) | accuracy: 43.61 -> 23.21 | Hurt |
| MAU-GPT 4B vs much larger 34B general models on MMAD | MAU-GPT 4B · MMAD (Average) | accuracy: 61.41 vs 67.16 (LLaVA-NeXT-34B) | Mixed |
Closest papers by meaning
- AnomalyGPT: Detecting Industrial Anomalies Using Large Vision-Language Models (AAAI 2024 · similarity 0.89)
- MMAD: A Comprehensive Benchmark for Multimodal Large Language Models in Industrial Anomaly Detection (ICLR 2025 · similarity 0.89)
- Towards Open-Vocabulary Industrial Defect Understanding with a Large-Scale Multimodal Dataset (CVPR 2026 · similarity 0.88)
- InCLAD: A Continual Learning Benchmark for Industrial Visual Anomaly Detection (NeurIPS 2026 · similarity 0.87)
- Supervised Anomaly Detection for Complex Industrial Images (CVPR 2024 · similarity 0.87)
- 3CAD: A Large-Scale Real-World 3C Product Dataset for Unsupervised Anomaly Detection (AAAI 2025 · similarity 0.87)
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.
