Apparatus, method, and program for supporting speech interface design
View Patent ↗For design of a speech interface accepting speech control options, speech samples are stored on a computer-readable medium. A similarity calculating unit calculates a certain indication of similarity of first and second sets of ones of the speech samples, the first set of speech samples being associated with a first speech control option and the second set of speech samples being associated with a second speech control option. A display unit displays the similarity indication. In another aspect, word vectors are generated for the respective speech sample sets, indicating frequencies of occurrence of respective words in the respective speech sample sets. The similarity calculating unit calculates the similarity indication responsive to the word vectors of the respective speech sample sets. In another aspect, a perplexity indication is calculated for respective speech sample sets responsive to language models for the respective speech sample sets.
1. A method using a computer, system having access to a database comprising speech samples, the speech samples comprising text data associated with a plurality of predetermined attributes, the method comprising steps of:
generating a first word vector for a first set of the speech samples, the first set being associated with a first attribute, and a second word vector for a second set of the speech samples, the second set being associated with a second attribute, wherein the first and second word vectors are generated based upon frequencies of occurrence of words in the speech samples contained in the first and second sets, respectively;
calculating a similarity between the first and second sets based upon the first and second word vectors;
using the computer system to generate, based upon the speech samples contained in the first and second sets, at least one language model for the first and second sets;
calculating a perplexity for each of the first and second sets, using the at least one language model for the first and second sets; and
displaying the similarity between the first and second sets and the perplexity for each of the first and second sets.
2. A method for supporting design of a speech interface of a system designed by a customer, the method using a computer system having access to a database comprising speech samples associated with at least one of a plurality of types of speech controls for the system designed by the customer, the method comprising:
generating a first word vector for a first set of the speech samples, the first set being associated with a first attribute, and a second word vector for a second set of the speech samples, the second set being associated with a second attribute, wherein the first and second word vectors are generated based upon frequencies of occurrence of words in the speech samples contained in the first and second sets, respectively;
calculating a similarity between the first and second sets based upon the first and second word vectors;
using the computer system to generate, based upon the speech samples contained in the first and second sets, at least one language model for the first and second sets;
calculating a perplexity for each of the first and second sets, using the at least one language model for the first and second sets;
displaying the similarity between the first and second sets and the perplexity for each of the first and second sets; and
generating an electronic report of results of analysis of the displayed similarity and perplexity.
3. The method according to claim 2 , further comprising:
transmitting the electronic report of the results of analysis to a computer of a customer via a network to which the computer is connected.
4. The method according to claim 2 , further comprising:
obtaining the speech samples for the plurality of types of speech controls; and
recording the speech samples to the database in association with at least one of the plurality of types of speech controls.