Scaffolded input promotes atomic organization in the recurrent neural network language model

Keyphrases

• Internal State 100%
• Lexical Representation 100%
• Recurrent Neural Network Language Model 100%
• Semantic Information 66%
• Training Data 66%
• Recurrent Neural Network 66%
• Controlled Experiment 33%
• Redundancy 33%
• Language Data 33%
• Lexical Item 33%
• Non-atomic 33%
• Artificial Language 33%
• Word Type 33%
• Sequential Dependencies 33%
• Redundant Information 33%
• Internal Organization 33%
• Language Applications 33%
• Semantic Categorization Task 33%
• Child-directed Language 33%
• Static Word Embeddings 33%
• Lexical Sequence 33%

Computer Science

• Language Modeling 100%
• Recurrent Neural Network 100%
• Internal State 75%
• Lexical Representation 75%
• Training Data 50%
• Application Language 25%
• Word Embedding 25%
• Internal Organization 25%