Deep Neural Networks Prove Promising In Modeling Human Hearing

In a new study from MIT, researchers have found that modern computational models derived from machine learning are moving closer to mimicking the structure and function of the human auditory system. The largest study yet of deep neural networks that have been trained to perform auditory tasks showed that most of these models generate internal representations that share properties of representations seen in the human brain when people are listening to the same sounds. The study also offers insight into how to best train this type of model, finding that models trained on auditory input including background noise more closely mimic the activation patterns of the human auditory cortex.