计算机科学
断层(地质)
班级(哲学)
蒸馏
人工智能
故障检测与隔离
机器学习
模式识别(心理学)
数据挖掘
地质学
执行机构
化学
地震学
有机化学
作者
Hanrong Zhang,Yifei Yao,Zixuan Wang,Jianwen Su,Mengxuan Li,Peng Peng,Hongwei Wang
标识
DOI:10.1109/tii.2025.3534407
摘要
Class-incremental fault diagnosis requires a model to adapt to new fault classes while retaining previous knowledge. However, limited research exists for imbalanced and long-tailed data. Extracting discriminative features from few-shot fault data is challenging, and adding new fault classes often demands costly model retraining. Moreover, incremental training of existing methods risks catastrophic forgetting, and severe class imbalance can bias the model's decisions toward normal classes. To tackle these issues, we introduce a supervised contrastive knowledge distillation for class incremental fault diagnosis (SCLIFD) framework proposing supervised contrastive knowledge distillation for improved representation learning capability and less forgetting, a novel prioritized exemplar selection method for sample replay to alleviate catastrophic forgetting, and the random forest classifier to address the class imbalance. Extensive experimentation on simulated and real-world industrial datasets across various imbalance ratios demonstrates the superiority of SCLIFD over existing approaches.
科研通智能强力驱动
Strongly Powered by AbleSci AI