In this work, we perform one of the first large-scale quantitative analysis of spontaneous aphasic speech based on automatic speech recognition (ASR) output.
"In this 10-year time frame, I believe that we'll not only be using the keyboard and the mouse to interact, but during that time we will have perfected speech recognition and speech output well enough that those will become a standard part of the interface".
In Nevisa, as in the most of the developed speech recognition systems, the output is a set of n-best hypotheses that are ordered based on their acoustic and language model scores.
Increasingly, people are using the Internet to handle their financial transactions, conduct job searches, plan vacations and take part in civic activities, yet many of the Web sites that offer these services are not designed to be accessible to users of voice output, speech recognition and other new technology.
The outputs of several speech recognition systems are first aligned, and a single word transcription network (WTN) is built.
Therefore, we retrained the DNN using the DAE output and performed speech recognition experiments.
We explore in this article a data-driven method where the broad classes are automatically defined according to the output of a speech recognition system.
This system looks for an exact match of the phone sequence output by a speech recognition process given a spoken query, within the phone lattices corresponding to the utterances.
In this paper, we propose a novel method to obtain discriminative feature transformation based on output coding technique for speech recognition.
There are two main approaches to STD: the word-based approach [6, 41 45] that searches for terms in the output of a large vocabulary continuous speech recognition (LVCSR) system, and the subword-based approach which searches for subword representations of search terms within the output of a subword speech recognition system.
