Structured semantic knowledge can emerge automatically from predicting word sequences in child-directed speech

Keyphrases

• Child-directed Speech 100%
• Semantic Knowledge 100%
• Word Sequence 100%
• Long Short-term Memory 80%
• Recurrent Neural Network 60%
• Abstract Knowledge 40%
• Learning Mechanisms 40%
• Distribution Learning 40%
• Structured Knowledge 40%
• Skip-gram 40%
• Hierarchical Structure 20%
• Child Age 20%
• Similarity Structure 20%
• Quantitative Evaluation 20%
• Learning Systems 20%
• Deep Learning Methods 20%
• Semantic Features 20%
• Grammatical Features 20%
• Performance Level 20%
• Distributed Representation 20%
• Statistical Learning 20%
• Word Corpus 20%
• Machine Learning 20%
• Internal Representation 20%
• Non-recurrent 20%
• 3-year-olds 20%
• Semantic System 20%
• Simple Recurrent Network 20%
• Sequential Dependencies 20%
• Taxonomic Relationships 20%
• Categorical Structure 20%
• Naturalistic Language 20%
• 0-3 Years 20%
• Naturalistic Speech 20%

Psychology

• Semantic Knowledge 100%
• Short-Term Memory 80%
• Neural Network 60%