Evidence of a predictive coding hierarchy in the human brain listening to speech

Scientists are studying how the human brain processes language compared to computer algorithms. They found that the way the human brain predicts what comes next in a story is different from how computers predict language. The human brain makes predictions that cover many parts of the story, while computers only predict the next few words. The scientists used brain scans to show that the brain predicts things hierarchically, meaning it considers larger parts of the story first before getting into smaller details. They also showed that modern computer language models can be improved by incorporating this hierarchical prediction method. This study helps us understand how the brain processes language and how computers can be improved to better mimic human language abilities.