Continual Learning of a Transformer-Based Deep Learning Classifier Using an Initial Model from Action Observation EEG Data to Online Motor Imagery Classification

Keyphrases

• EEG Data 100%
• Motor Imagery 100%
• Action Observation 100%
• Deep Learning Classifier 100%
• Continual Learning 100%
• Motor Imagery Classification 100%
• Transformer-based Deep Learning 100%
• Brain-computer Interface 30%
• Motor Imagery Task 30%
• Transformer-based 23%
• Spatial-temporal Network 23%
• Virtual Environment 15%
• Neural Network Method 7%
• EEG Signals 7%
• Big Data 7%
• Convolutional Neural Network 7%
• Attention Mechanism 7%
• Detection Accuracy 7%
• Spatial Features 7%
• Auditory Cues 7%
• External Environment 7%
• Spatial Dimension 7%
• Recurrent Neural Network 7%
• Training Methods 7%
• Nave 7%
• Training Process 7%
• Temporal Features 7%
• Interface Performance 7%
• Brain-computer Interface Systems 7%
• Temporal Dimension 7%
• Training Session 7%
• Data Training 7%
• Motor Imagery Brain-computer Interface 7%
• Visual Feedback 7%
• Audiovisual Integration 7%
• Cross-subject 7%
• Head-mounted Device 7%
• EEG Pattern 7%
• Global Dependencies 7%
• Inter-subject Variability 7%
• Intuitive Interfaces 7%

Computer Science

• Computer Interface 100%
• Transformer-Based Deep Learning 100%
• Virtual Reality 40%
• Neural Network Approach 20%
• Big Data 20%
• Convolutional Neural Network 20%
• Attention (Machine Learning) 20%
• Detection Accuracy 20%
• Training Process 20%
• Interface System 20%
• Temporal Dimension 20%
• Training Session 20%
• Recurrent Neural Network 20%
• Spatial Dimension 20%
• Healthy Subject 20%
• Temporal Feature 20%
• Subject Variability 20%
• Visual Feedback 20%
• External Environment 20%