Systems and Methods for Tuning Automatic Speech Recognition Systems

Abstract

A tuning system for tuning a speech recognition system includes a transmitter for sending a user response to a speech recognition system. The user response is based at least in part on a test stimulus that may be generated by the control system. A receiver receives a recognized response from the speech recognition system; this recognized response is based at least in part on the associated user response. An adjustment module adjusts at least one parameter of the speech recognition system based at least in part on at least one of the test stimulus, the associated user response, and the recognized response.

Description

CROSS-REFERENCE TO RELATED APPLICATIONS[0001]This application claims the benefit of U.S. Provisional Patent Application No. 61 / 164,451, filed Mar. 29, 2009, the disclosure of which is hereby incorporated by reference herein in its entirety.FIELD OF THE INVENTION[0002]This invention relates generally to tuning automatic speech recognition systems, and more specifically, self-tuning automatic speech recognition systems based on a user's speech model.BACKGROUND[0003]Multi-channel Cochlear Implant (CI) systems consist of an external headset with a microphone and transmitter, a body-worn or ear-level speech processor with a battery supply, and an internal receiver and electrode array. The microphone detects sound information and sends it to the speech processor which encodes the sound information into a digital signal. This information then is sent to the headset so that the transmitter can send the electrical signal through the skin via radio frequency waves to the internal receiver loc...

AI Technical Summary

Benefits of technology

[0020]In another embodiment, the method can include adjusting at least one of the operational parameters of the communication channel according to the comparing and associating steps. Notably, the communication channel can include an acoustic environment formed by an architectural structure, an underwater acoustic environment, or the communication channel can mimic aviation effects on speech and hearing. For example, the communication channel can mimic effects such as G-force, masks, and the Lombard effect on hearing. The steps disclosed herein also can be performed in cases where the user exhibits signs of stress or fatigue.

Problems solved by technology

The tuning process, however, is not a user-specific process.

This approach has gained favor because the amount of time and the number of potential variables involved in designing optimal maps for individual users would be too daunting a task.

For example, additional complications to the tuning process exist when users attempt to add subjective input to the tuning of the hearing enhancement system.

Using subjective input from a user can add greater complexity to the tuning process as each change in the mapping of a hearing enhancement system requires the user to adjust to a new signal.

However, if that ASR system is being used by a person with a speech impediment, it may incorrectly recognize certain stimuli, and may dial the incorrect contact or take other inappropriate action.

Research has been performed regarding such systems for tuning ASR systems, but the resulting tuning systems still display performance limitations.

Images