New framework CoEvoAD enhances zero-shot anomaly detection with evolutionary prompt selection

Researchers have developed CoEvoAD, a novel co-evolutionary framework for zero-shot anomaly detection that operates in the discrete natural-language space. This approach uses an evolutionary algorithm to search for and select prompts, preserving interpretability and composability. To enhance generalization across categories, a Cross-Category Transfer Objective (CCTO) was introduced, which estimates prompt transferability to unseen categories. Experiments demonstrate that CoEvoAD achieves state-of-the-art performance on various anomaly detection datasets. AI

IMPACT This new method could improve the accuracy and interpretability of anomaly detection systems in industrial applications.

RANK_REASON The cluster contains a research paper detailing a new method for anomaly detection. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.CV →

• alphaXiv
• arXiv
• CatalyzeX
• CoEvoAD
• Cross-Category Transfer Objective
• DagsHub
• Gotit.pub
• Hugging Face
• ScienceCast
• Zero-shot anomaly detection

• paper
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →