<ul><li>The paper introduces a novel method called SPLAENet for Stance Prediction in misinformative social media content.</li><li>SPLAENet utilizes a dual cross-attention mechanism, hierarchical attention network, and emotions to distinguish between different stance categories in user-generated content.</li><li>The method incorporates label fusion using distance-metric learning to align extracted features with stance labels, enhancing accuracy in distinguishing stances.</li><li>Experiments show that SPLAENet outperforms existing methods with significant improvements in accuracy and F1-score across different datasets, validating its effectiveness in stance detection.</li></ul>

Emotion-aware Dual Cross-Attentive Neural Network with Label Fusion for Stance Detection in Misinformative Social Media Content

Discover more