Multi-view domain-adaptive representation learning for EEG-based emotion recognition

Chao Li, Ning Bian, Ziping Zhao, Haishuai Wang, Björn W. Schuller

Abstract: Highlights•Our method can extract discriminative features and minimize individual differences.•Multi-view learning explores the complementarity between different channels.•Multi-view cross-attention extracts discriminative spatio-temporal features.•Dilated causal convolutional neural network captures causal interactions among EEG.•Domain discriminator makes feature distributions similar for better generalization.