Voice input system for data retrieval

• US 5,454,063 A
• Filed: 11/29/1993
• Issued: 09/26/1995
• Est. Priority Date: 11/29/1993
• Status: Expired due to Fees

• Current Assignee: Michael T. Rossides
• Original Assignee: Michael T. Rossides
• Inventors: Rossides, Michael T.
• Primary Examiner(s): MacDonald, Allen R.
• Assistant Examiner(s): ONKA, THOMAS
• Application Number:
  US08/158,297
• Time in Patent Office: 666 Days
• Field of Search: 364/419.03, 364/419.04, 364/419.15, 395/2.69, 395/2.63, 395/2.79, 395/2.84
• US Class Current: 704/275
• CPC Class Codes: G06F 16/90348 by searching ordered data, e.g. alpha-numerically ordered data
• CPC Class Codes: G10L 15/26 Speech to text systems G10L15/08 takes precedence

Abstract

A computer input system employing an automatic speech recognizer. The system is used for finding names in a data-base. The system reduces the number of letters a speaker needs to enter in order to find a name uniquely in the data-base. The system enables the speaker to enter inputs that identify which words in the name the speaker'"'"'s letter inputs correspond to. The system builds a set of search parameters incorporating both the user'"'"'s word identifier inputs and letter inputs. This set can be called an abbreviation because it usually represents a small fraction of the total number of letters in the name the user seeks to find.

Claims

1. An input system employing an automatic speech recognizer that can recognize and confirm words and numbers used as program commands and can recognize and confirm alpha-numeric symbols (the letters of the alphabet and the digits 0-9);

said input system being used for entering search parameters into a data-base system, in order to find names in said data-base;

said input system builds a set of search parameters, called an abbreviation, said abbreviation minimizing the inputs necessary to specify a name in said data-base;

said input system comprising in combination the following elements;

a computer with memory, processor, and output means, along with said speech recognizer as an input means, and a program directing the operation of said input system;

said program enabling the system to accept speech inputs from a speaker speaking into said recognizer, said program enabling the system to distinguish among three sets of inputs;

a. letter inputs that make up words, said letter inputs including the letters of the alphabet and the digits 0-9,

b. word identifier inputs that denote the ordinal position of words in a name, each identifier including a different number corresponding to a different ordinal position for a word, each number being used to define a field that identifies which word a speaker'"'"'s letter inputs correspond to,

c. a termination input that signifies that no more letters or word identifiers will be stored in an abbreviation;

said program directing said input system to execute the following steps;

first, waiting for a word identifier to be entered, then, after the entry of said identifier, beginning to build an abbreviation and, storing said word identifier in said abbreviation, after which,

if a word identifier is entered, storing it in said abbreviation,

if a letter is entered, storing it in said abbreviation in the field defined by the last word identifier entered,

if a termination input is entered, storing no more inputs in said abbreviation and returning to the first step above;

said input system being combined with a data-base system provided with search means that use said abbreviation to find a name and data corresponding to said name in said data-base;

said data-base system executing the following steps upon an input being stored in said abbreviation;

searching said data-base for a name uniquely matching said abbreviation,

if said abbreviation matches no name, outputting a message saying that no name matches the name being abbreviated,

if said abbreviation matches a set of multiple names, outputting no message and waiting for another input to be stored in said abbreviation,

if said abbreviation matches one name, outputting said name and data corresponding to said name.

2. The system of claim 1 wherein said program further distinguishes another set of inputs:

said inputs being letter identifier inputs that denote the ordinal position of letters in a word, each identifier including a different number corresponding to a different ordinal position, each number being used to define a field that identifies which position in a word a letter input corresponds to,

said program further directing said input system to execute the following steps, if a letter identifier is entered,

storing said letter identifier in said abbreviation in the field defined by the last word identifier entered and,

if the next input entered is a letter, storing said letter in the field defined by said letter identifier.

3. The system of claim 1 wherein said program further distinguishes another set of inputs:

said inputs being numeric value inputs that denote the number of letters in the word corresponding to the last word identifier entered, each numeric value input including a different number for denoting a different number of letters in a word,

said program further directing said input system to execute the following steps, if a numeric value input is entered,

storing the number of said numeric value input in said abbreviation in the field defined by the last word identifier entered and,

using said number to denote the number of letters in the word corresponding to said word identifier.

4. The system of claim 1 wherein said data-base system is provided with calculation means for calculating an expected information value of letter inputs, said means calculating the percentage of false matches, on average, that inputs eliminate from a set of names being searched,

said step, if said abbreviation matches a set of multiple names, further comprising,

checking if the number of letters stored in said abbreviation is above a threshold,

if no, outputting no message and waiting for another input to be stored in said abbreviation,

if yes, taking the set of matched names,

from this set, taking the set of all the words corresponding to the last word identifier entered,

setting a string length, N,

calculating the expected value of the next N letters in all the words taken,

if the expected value is below a threshold, outputting a message that little progress will be made by spelling the word corresponding to the last word identifier entered,

if the expected value is not below the threshold, outputting no message and waiting for another input to be stored in said abbreviation.

5. The system of claim 1 wherein said data-base system is provided with calculation means for calculating an expected information value word identifier and letter inputs, and further provided with a stored set of word identifier and letter inputs called suggested inputs,

said means calculating the percentage of false matches, on average, that inputs eliminate from a set of names being searched,

said step, if said abbreviation matches a set of multiple names, further comprising,

calculating the expected information value of each input in the set of suggested inputs,

finding the suggested input with the highest expected value,

checking if said highest expected value is above a threshold,

if no, outputting no message and waiting for another input to be stored in said abbreviation,

if yes, outputting a message suggesting the entering of the input corresponding to said highest expected value.