Evidence of a predictive coding hierarchy in the human brain listening to speech
École Polytechnique · Commissariat à l'Énergie Atomique et aux Énergies Alternatives · +6 more institutions
Abstract
Considerable progress has recently been made in natural language processing: deep learning algorithms are increasingly able to generate, summarize, translate and classify texts. Yet, these language models still fail to match the language abilities of humans. Predictive coding theory offers a tentative explanation to this discrepancy: while language models are optimized to predict nearby words, the human brain would continuously predict a hierarchy of representations that spans multiple timescales. To test this hypothesis, we analysed the functional magnetic resonance imaging brain signals of 304 participants listening to short stories. First, we confirmed that the activations of modern language models linearly…
Citation impact
- FWCI
- 48.15
- Percentile
- 100%
- References
- 85
Authors
3- CCCharlotte CaucheteuxCorresponding
École Polytechnique, Commissariat à l'Énergie Atomique et aux Énergies Alternatives, Université Paris-Saclay, Milieux environnementaux, transferts et interactions dans les hydrosystèmes et les sols
- AGAlexandre Gramfort
École Polytechnique, Commissariat à l'Énergie Atomique et aux Énergies Alternatives, Université Paris-Saclay
- JKJean-Rémi King
Centre National de la Recherche Scientifique, Université Paris Sciences et Lettres, École Normale Supérieure - PSL, Milieux environnementaux, transferts et interactions dans les hydrosystèmes et les sols, Laboratoire des Systèmes Perceptifs
Topics & keywords
- Active listening
- Predictive coding
- Hierarchy
- Functional magnetic resonance imaging
- Coding (social sciences)
- Computer science
- Cognition
- Language model
- Quality Education