Browse Prior Art Database

High-speed Context Switching within Tree Structures in a Speech Recognition System

IP.com Disclosure Number: IPCOM000114588D
Original Publication Date: 1995-Jan-01
Included in the Prior Art Database: 2005-Mar-29
Document File: 2 page(s) / 37K

Publishing Venue

IBM

Related People

Castellucci, F: AUTHOR [+2]

Abstract

Disclosed is a speech recognition system in which data elements, organized into sets with typically fewer than twenty to fifty items per set, are assigned one or more acoustic signature(s) or phoneme string(s) and a matching text string, which can be displayed on the screen. If the number of items is fewer than the desired limit of twenty to fifty, the organization process stops at this point; otherwise the process is repeated with various sets grouped into mega-sets.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 86% of the total text.

High-speed Context Switching within Tree Structures in a Speech Recognition
System

      Disclosed is a speech recognition system in which data
elements, organized into sets with typically fewer than twenty to
fifty items per set, are assigned one or more acoustic signature(s)
or phoneme string(s) and a matching text string, which can be
displayed on the screen.  If the number of items is fewer than the
desired limit of twenty to fifty, the organization process stops at
this point; otherwise the process is repeated with various sets
grouped into mega-sets.

      From the point of view of the user, a list of topics,
categories, or questions is presented on the screen.  Each item in
the list has a set name or a mega-set name.  For example, in a
hospital system, the user might be asked to choose among "Getting
ready for your baby," "The delivery," "The first month," "One year
olds," etc.  If "One year olds" is a mega-set, the user making this
choice may be presented with a menu of topics, and, under that, the
most common questions asked.

      In this way, a tree structure is created, including a table of
contents, or a super table of contents providing entries into various
tables of contents.  This approach uses the ability of
high-performance
speech recognition systems to employ high-speed context switching
within
tree structures, quickly changing the active vocabularies, BNFs
(Backus-Naur Forms), and grammars used by the system.  Users can
quickly
scan t...