Does syntax need to grow on trees? Sources of hierarchical inductive bias in sequence-to-sequence networks

Computer Science

• Syntactics 100%
• Training Data 33%
• Neural Network Model 33%
• Generalized Model 33%
• Model Architecture 33%
• Long Short-Term Memory Network 33%
• Gated Recurrent Unit 33%
• Structured Model 33%
• Sequence Model 33%

Psychology

• Neural Network 100%
• Network Model 100%
• Training Set 100%

Keyphrases

• Sequence Networks 100%
• Syntactic Generalization 25%
• Generalization Behavior 25%