Research map / Vision-language and multimodal / CLIP and zero-shot
PromptMoE: Generalizable Zero-Shot Anomaly Detection via Visually-Guided Prompt Mixtures
Yuheng Shao, Lizhang Wang, Changhao Li, Peixian Chen, Qinyuan Liu
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 |
|---|---|---|---|
| replace single static prompt with static ensemble (ablation) | PromptMoE · MVTec AD | I-AUC: 91.7 -> 92.2 | Improved |
| replace static ensemble with single-layer VGMoP (ablation) | PromptMoE · MVTec AD | I-AUC: 92.2 -> 93.1 | Improved |
| replace static ensemble with single-layer VGMoP (ablation) | PromptMoE · VisA | I-AUC: 83.3 -> 84.1 | Improved |
| add multi-layer prompts and auxiliary losses to VGMoP (ablation) | PromptMoE · VisA | I-AUC: 84.1 -> 85.0 | Improved |
| add multi-layer prompts and auxiliary losses to VGMoP (ablation) | PromptMoE · MVTec AD | PRO: 83.3 -> 83.2 | No clear effect |
| add load-balancing loss (alpha 0 -> 0.01) (ablation) | PromptMoE · MVTec AD | I-AUC: 92.1 -> 93.8 | Improved |
| increase load-balancing weight from 0.01 to 0.1 (ablation) | PromptMoE · MVTec AD | I-AUC: 93.8 -> 92.9 | Hurt |
| add expert decoupling loss (ablation) | PromptMoE · MVTec AD | I-AUROC | Improved |
Closest papers by meaning
- VarProtoAD: Variational Prototype-Conditioned Prompting for Zero-Shot Anomaly Detection (ECCV 2026 · similarity 0.93)
- Exploring High-order-aware Prompt Learning for Zero-shot Anomaly Detection (AAAI 2026 · similarity 0.93)
- AnomalyCLIP: Object-agnostic Prompt Learning for Zero-shot Anomaly Detection (ICLR 2024 · similarity 0.91)
- Bayesian Prompt Flow Learning for Zero-Shot Anomaly Detection (CVPR 2025 · similarity 0.91)
- PromptAD: Learning Prompts with only Normal Samples for Few-Shot Anomaly Detection (CVPR 2024 · similarity 0.91)
- Towards Zero-Shot Anomaly Detection and Reasoning with Multimodal Large Language Models (CVPR 2025 · similarity 0.90)
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.
