Deep neural networks and humans both benefit from compositional language structure
Uloženo v:
| Název: | Deep neural networks and humans both benefit from compositional language structure |
|---|---|
| Autoři: | Lukas Galke, Yoav Ram, Limor Raviv |
| Zdroj: | Nat Commun Nature Communications, Vol 15, Iss 1, Pp 1-13 (2024) |
| Informace o vydavateli: | Springer Science and Business Media LLC, 2024. |
| Rok vydání: | 2024 |
| Témata: | 0301 basic medicine, 0303 health sciences, Neural Networks, Science, Learning/physiology, Linguistics, Article, Computer, 03 medical and health sciences, Deep Learning, Humans, Learning, Neural Networks, Computer, Language, Natural Language Processing |
| Popis: | 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. |
| Druh dokumentu: | Article Other literature type |
| Jazyk: | English |
| ISSN: | 2041-1723 |
| DOI: | 10.1038/s41467-024-55158-1 |
| Přístupová URL adresa: | https://pubmed.ncbi.nlm.nih.gov/39738033 https://doaj.org/article/fae43946b97447cf9d3969e4a7ae2cc3 |
| Rights: | CC BY |
| Přístupové číslo: | edsair.doi.dedup.....b1358f95b001f4e5d36bb7a1a1ef779e |
| Databáze: | OpenAIRE |
Full Text 
Full Text Finder 
Nájsť tento článok vo Web of Science | Abstrakt: | 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. |
|---|---|
| ISSN: | 20411723 |
| DOI: | 10.1038/s41467-024-55158-1 |