Metric-based meta-learning model for few-shot fault diagnosis under multiple limited data conditions

Keyphrases

• Meta-learning Model 100%
• Limited Data Condition 100%
• Few-shot Fault Diagnosis 100%
• Metric Learning 100%
• Feature Space 25%
• Learning Task 25%
• Supervised Learning 25%
• Information Technology 25%
• Sensor Technology 25%
• T-distributed Stochastic Neighbor Embedding (t-SNE) 25%
• Implementation Problems 25%
• Time Complexity 25%
• Few Samples 25%
• Experiment Results 25%
• Catastrophic Failure 25%
• Fault Type 25%
• Data Environment 25%
• Large-scale Industry 25%
• Attribute Information 25%
• Bearing Fault Diagnosis 25%
• Gear Fault Diagnosis 25%
• Failure Attribution 25%
• Data-driven Fault Diagnosis 25%
• Feature Embedding 25%
• Meta-learning 25%
• Labeled Sample 25%
• Similarity Information 25%

Computer Science

• Meta-Learning 100%

Chemical Engineering

• Supervised Learning 100%