Automatic Speech Recognition on Par with Humans in Noisy Conditions
Current ASR systems achieve remarkable accuracy and may even surpass human performance. While they need to be trained on an incredible amount of data, humans acquire comparable skills in less time.