Machine interface

• US 6,832,197 B2
• Filed: 03/13/2001
• Issued: 12/14/2004
• Est. Priority Date: 03/23/2000
• Status: Expired due to Fees

• Current Assignee: Canon Kabushiki Kaisha (Canon Inc.)
• Original Assignee: Canon Kabushiki Kaisha (Canon Inc.)
• Inventors: Hogenhout, Wide Roeland
• Primary Examiner(s): ABEBE, DANIEL DEMELASH
• Application Number:
  US09/803,970
• Publication Number:
  US 20020004719A1
• Time in Patent Office: 1,372 Days
• Field of Search: 704/272, 704/275, 704/240, 704/207, 704/257, 704/260, 379/88.01
• US Class Current: 704/275
• CPC Class Codes: G06F 16/3329 Natural language query formulation or dialogue systems
• CPC Class Codes: G10L 13/00 Speech synthesis; Text to speech systems
• CPC Class Codes: G10L 15/22 Procedures used during a speech recognition process, e.g. man-machine dialogue

Abstract

A machine interface for interfacing a user to a machine receives input words from a user. Scores for at least some of the input words are determined using scores for words appearing in each passages of the text. A passage in the text having the highest combined score for the input words is identified and a corresponding dialogue instruction set is executed to generate words for output and to respond to input words from the user.

Claims

1. A machine having a user interface comprising:

inputting means for inputting words from a user;

outputting means for outputting words to a user;

text inputting means for inputting a text document;

fragmentation means for identifying a plurality of passages in said text document;

instruction set generation means for generating, for each one of said plurality of passages, a corresponding dialogue instruction set to provide a plurality of said dialogue instruction sets with each dialogue instruction set corresponding to a respective different one of said plurality of passages;

score determining means for determining scores for words appearing in each of said plurality of passages;

instruction set storage means for storing said plurality of dialogue instruction sets;

score storage means for storing said scores;

search means for receiving the input words, for determining scores for at least some of the input words from said score storage means, and for identifying the one of said plurality of passages in said text having a highest combined score for the input words; and

dialogue execution means for executing a dialogue instruction set corresponding to said passage identified as having the highest combined score to generate words for output by said outputting means to the user in response to words input by said inputting means.

2. A machine according to claim 1, further comprising user prompt means for generating a user prompt corresponding to said passage having the highest combined score for output by said outputting means if said search means does not determine a said passage with a combined score significantly higher than the combined scores for other said passages, said dialogue execution means being responsive to one or more positive input words to execute said dialogue instruction set, and said search means being responsive to one or more negative input words to reduce the score for said passage and to redetermine a said passage having a highest combined score.

3. A machine according to claim 1 for providing a spoken dialogue interface, wherein said inputting means is adapted to input spoken words, said outputting means is adapted to output spoken words, said instruction set storage means is adapted to store spoken dialogue instruction sets, and said dialogue execution means is adapted to execute a spoken dialogue instruction set.

4. A machine according to claim 1, wherein said search means is adapted to ignore non information bearing words when determining the scores.

5. Configuration apparatus for generating dialogue configuration data, the apparatus comprising:

text inputting means for inputting a text document;

fragmentation means for identifying a plurality of passages in said text document;

instruction set generation means for generating, for each one of said plurality of passages, a corresponding dialogue instruction set to provide a plurality of said dialogue instruction sets with each dialogue instruction set corresponding to a respective different one of said plurality of passages and for storing said dialogue instruction sets in instruction set storage means; and

score determining means for determining scores for words appearing in each of said plurality of passages and for storing said scores in score storage means.

6. An interface method for a machine, the method comprising:

receiving a text document;

identifying a plurality of passages in said text document;

generating, for each one of said plurality of passages, a corresponding dialogue instruction set to provide a plurality of said dialogue instruction sets with each dialogue instruction set corresponding to a respective different one of said plurality of passages;

storing said dialogue instruction sets;

determining scores for words appearing in each of said plurality of passages;

storing said scores;

receiving words from a user;

determining scores for at least some of the words received from the user using said stored scores;

identifying the one of said plurality of passages in said text document having a highest combined score for the words received from the user; and

identifying and executing the one of said plurality of dialogue instruction sets corresponding to said identified passage to generate words for output to the user in response to the received words.

7. Processor implementable instruction code for controlling a processor to implement the method of claim 6.

8. A storage medium storing the processor implementable instruction code according to claim 7.

9. An interface method according to claim 6, further comprising:

a) generating a user prompt corresponding to said identified passage if said identified passage does not have a combined score significantly higher than the combined scores for other said passages,

b) if a positive response is received from the user, executing said dialogue instruction set or if a negative response is received from the user, reducing the score for said identified passage and re-identifying a said passage having a highest combined score, and

c) repeating steps a) and b) if a said dialogue instruction set is not executed.

10. An interface method according to claim 9, wherein if anything other than a positive or negative response is received from the user, the steps of claim 6 are repeated using the words of the response as the received words from the user.

11. An interface method according to claim 6, for providing a spoken dialogue interface for the machine, wherein the words received from the user are spoken words, said dialogue instruction set comprises a spoken dialogue instruction set which is executed to generate spoken words for output and to respond to received spoken words.

12. An interface method according to claim 6, wherein the step of determining scores comprises determining scores for only information bearing received words.

13. A method of generating dialogue configuration data, the method comprising the steps of:

receiving a text document;

identifying a plurality of passages in said text document;

generating, for each one of said plurality of passages, a corresponding dialogue instruction set to provide a plurality of said dialogue instruction sets with each dialogue instruction set corresponding to a respective different one of said plurality of passages;

storing said plurality of dialogue instruction sets;

determining scores for words appearing in each of said plurality of passages; and

storing said scores.

14. A storage medium storing said dialogue instruction sets and said scores for words generated in accordance with claim 13.

15. Configuration apparatus for generating dialogue configuration data, the apparatus comprising:

a text receiver operable to receive a text document;

a fragmenter operable to identify a plurality of passages in said text document;

an instruction set generator operable to generate, for each one of said plurality of passages, a corresponding dialogue instruction set to provide a plurality of said dialogue instruction sets with each dialogue instruction set corresponding to a respective different one of said plurality of passages and to store said dialogue instruction sets in an instruction set store; and

a score determiner operable to determine scores for words appearing in each passage and to store said scores in a score store.

16. Apparatus for conducting a dialogue with a user, the apparatus comprising:

configuration apparatus in accordance with claim 15;

a user input receiver operable to receive words from a user;

an output provider operable to output words to a user;

a passage identifier operable to determine from said score store scores for at least some of the words received by the user input receiver and to identify the one of said plurality of passages having a highest combined score for the received words; and

a dialogue executor operable to execute the dialogue instructions set corresponding to said identified passage to generate words for output to the user by said output provider means in response to the received words.

17. Configuration apparatus for generating dialogue configuration data from text data containing mark-up language tags, the apparatus comprising:

a text receiver operable to receive text data containing mark-up language tags;

a fragmenter operable to define text fragments in the text data on the basis of the mark-up language tags in the text data to provide a plurality of text fragments;

a dialogue instruction set generator operable to generate, for each one of said plurality of text fragments, a corresponding dialogue instruction set to provide a plurality of said dialogue instruction sets with each dialogue instruction set corresponding to a respective different one of said plurality of text fragments; and

a score determiner operable to determine scores for words appearing in each of said plurality of text fragments.

18. Apparatus for conducting a dialogue with a user, the apparatus comprising:

configuration apparatus in accordance with claim 17;

a user input receiver operable to receive words from a user;

an output provider operable to output words to a user;

a passage identifier operable to determine from said score store scores for at least some of the words received by the user input receiver and to identify the one of said plurality of passages having a highest combined score for the received words; and

a dialogue executor operable to execute the dialogue instruction set corresponding to said identified passage to generate words for output to the user by said output provider means in response to the received words.

19. A method of generating dialogue configuration data from text data containing mark-up language tags, the method comprising the steps of:

receiving text data containing mark-up language tags;

defining text fragments in the text data on the basis of the mark-up language tags in the text data to provide a plurality of text fragments;

generating, for each one of said plurality of text fragments, a corresponding dialogue instruction set to provide a plurality of said dialogue instruction sets with each dialogue instruction set corresponding to a respective different one of said plurality of text fragments; and

determining scores for words appearing in each of said plurality of text fragments.

20. A storage medium storing program instructions for programming a processor to carry out a method in accordance with claim 19.