Sign In

Communications of the ACM

ACM TechNews

Speech Recognition Systems Must Get Smarter, Professor Says

View as: Print Mobile App Share:
University of Rochester professor James F. Allen

Great strides have been made in voice recognition and natural language processing over the past few decades, but they have seemingly brought mostly frustration to users, says University of Rochester professor James F. Allen.

Credit: University of Rochester

Most modern computerized speech-recognition systems can understand what a human says up to 98 percent of the time, yet people still get frustrated using automated phone help-desk systems, says University of Rochester professor James Allen. He says the key to making speech-recognition systems less frustrating to use is giving them a deeper understanding of language and making them more interactive. Allen has been researching ways to make these systems more life-like in the way they interact with humans. The goal is to be able to "talk to a machine the same way we can talk to a person," he says.

A program designed by Allen and his team, called Plow, can learn how to carry out common tasks on a computer. "This is a system that allows you to essentially use dialog to train your system how to do things for you," he says.

Another program designed by Allen and his research team, called Cardiac, mimics the questions a nurse asks a patient with heart disease. The system determines what information was provided and what is still needed. However, Allen says better two-way communications between users and computers is still needed.

From PC World
View Full Article


Abstracts Copyright © 2010 Information Inc., Bethesda, Maryland, USA


No entries found

Sign In for Full Access
» Forgot Password? » Create an ACM Web Account