EDDA: An Encoder-Decoder Data Augmentation Framework for Zero-Shot Stance Detection

Keyphrases

• Data Augmentation 100%
• Encoder-decoder 100%
• Zero-shot Stance Detection 100%
• Training Data 20%
• Benchmark Dataset 20%
• Detection Method 20%
• Knowledge Transfer 20%
• Sampling Methods 20%
• Unseen 20%
• Source Text 20%
• Decoder 20%
• Encoder 20%
• Replacement Strategy 20%
• Target Text 20%
• Logical Connection 20%
• Logical Relationship 20%
• Semantic Relevance 20%
• Stance Detection 20%
• Augmented Data 20%
• Large Language Models 20%
• Chain-of-Thought 20%
• Semantic Correlation 20%
• Enhanced Network 20%
• Syntactic Diversity 20%
• Word Replacement 20%
• Text Augmentation 20%
• If-then-else 20%

Computer Science

• Data Augmentation 100%
• Stance Detection 100%
• Zero-Shot Learning 100%
• Syntactics 40%
• Training Data 20%
• Logical Relationship 20%
• Large Language Model 20%
• Semantic Relevance 20%
• Chain-of-Thought 20%

Engineering

• Data Augmentation 100%
• Limitations 20%
• Detection Technique 20%
• Large Language Model 20%