What makes a language easy to deep-learn? Deep neural networks and humans similarly benefit from compositional structure
This addresses the problem of understanding language learnability for AI models, with incremental implications for using neural networks as models of human language learning.
The study tested whether deep neural networks, like humans, find languages with more compositional structure easier to learn and generalize, showing that neural networks exposed to such languages exhibit more systematic generalization, greater agreement between agents, and greater similarity to human learners.
Deep neural networks drive the success of natural language processing. A fundamental property of language is its compositional structure, allowing humans to systematically produce forms for new meanings. For humans, languages with more compositional and transparent structures are typically easier to learn than those with opaque and irregular structures. However, this learnability advantage has not yet been shown for deep neural networks, limiting their use as models for human language learning. Here, we directly test how neural networks compare to humans in learning and generalizing different languages that vary in their degree of compositional structure. We evaluate the memorization and generalization capabilities of a large language model and recurrent neural networks, and show that both deep neural networks exhibit a learnability advantage for more structured linguistic input: neural networks exposed to more compositional languages show more systematic generalization, greater agreement between different agents, and greater similarity to human learners.