Method and system for automatically verbally responding to user inquiries about information

• US 6,647,363 B2
• Filed: 10/07/1999
• Issued: 11/11/2003
• Est. Priority Date: 10/09/1998
• Status: Expired due to Term

• Current Assignee: Nuance Communications, Inc. (Microsoft Corporation)
• Original Assignee: ScanSoft, Inc. n/k/a Nuance Communications, Inc. (Microsoft Corporation)
• Inventors: Claassen, Antonius M. W.
• Primary Examiner(s): EDOUARD, PATRICK NESTOR
• Application Number:
  US09/414,459
• Publication Number:
  US 20030125926A1
• Time in Patent Office: 1,496 Days
• Field of Search: 704/1, 704/9, 704/10, 704/270, 704/272, 704/275, 705/26, 379/88.01, 379/88.04, 379/88.16, 379/88.18
• US Class Current: 704/1
• CPC Class Codes: G06F 16/24522 Translation of natural language queries to structured queries
• CPC Class Codes: G06F 16/3329 Natural language query formulation or dialogue systems
• CPC Class Codes: G10L 15/1815 Semantic context, e.g. disambiguation of the recognition hypotheses based on word meaning
• CPC Class Codes: G10L 15/22 Procedures used during a speech recognition process, e.g. man-machine dialogue

Abstract

A system is presented for automatically responding to a user inquiry comprising a dialog manager and a presentation manager. The dialog manager executes a machine-controlled human/machine dialog to determine a set of query items, and in response thereto, retrieves information items from memory. The presentation manager determines the inquiring user'"'"'s intentions motivating and associated with the query, and in response thereto selects a preferred manner of presenting the retrieved information items, or presentation scenario. In so doing, at least one natural language phrase is generated to match the selected presentation scenario, and a speech generator verbally presents the generated phrasing to the inquiring user.

Claims

1. A method for automatically responding to an inquiry from a user, comprising:

executing a machine-controlled human-machine dialogue to determine a plurality of predetermined query items specifying information to be verbally presented to the user;

retrieving a plurality of information items from a storage in dependence on the query items;

determining an intention of the user from a predetermined set of intentions, the intention reflecting a preferred way of presenting the information items;

selecting the verbal presentation scenario from a predetermined set of presentation scenarios in dependence on the determined intention, wherein the presentation scenarios are selected from the group consisting of semantic content of delivered information items, emphasis on or repetition of various information items, and tailoring of actual vocabulary used;

generating at least one natural language phrase to present the obtained information items according to a presentation scenario; and

verbally presenting the generated phrase(s) to the user.

2. The method of claim 1, wherein the determination of an intention comprises determining which of the information items is/are important to the user.

3. The method of claim 2, further comprising nominating an information item as important if a corresponding query item occurs early in a user utterance relative to other query items.

4. The method of claim 2, wherein selection of the presentation scenario comprises selecting from the predetermined set of presentation scenarios a scenario wherein the information item(s) nominated as important is/are presented early in the phrase(s) relative to the other information items.

5. The method of claim 1, wherein the execution of the dialogue comprises:

outputting a question and/or a verifying statement from a statement repertoire to the user; and

obtaining at least one response from the user;

said execution being continued until for a plurality of pre-determined query specifiers a respective query item has been recognized in the response(s); and

wherein the determination of an intention comprises determining the intention from the response(s) of the user to a question and/or verifier statement.

6. The method of claim 5, where:

each of the intentions of the predetermined set of intentions is associated with a respective intention specifier;

at least one of the intention specifiers does not correspond to one of the query specifiers; and

said method further comprising trying to recognize an intention for at least one of the intention specifiers in the response(s) during the dialogue.

7. The method of claim 6, further comprising selecting a default presentation scenario if no intention item has been recognized during the dialogue.

8. The method of claim 1, further comprising:

detecting that the user barges in during presentation of the generated phrases; and

determining an intention of the user from a moment of the user barging in and/or a user barging-in utterance;

re-selecting a presentation scenario based on the intention of the user motivating his barging-in;

generating at least one phrase representing at least the not yet presented information items according to the selected presentation scenario; and

verbally presenting the generated phrase(s) to the user.

9. The method of claim 1, wherein the method comprises:

comparing the complexity of a phrase representing the obtained information items according to a default or selected presentation scenario with a predetermined threshold; and

if judged too complex,

outputting a question statement to the user for determining a more suitable presentation scenario; and

determining the intention of the user from a reply to the question statement.

10. The method of claim 1, wherein the set of intentions includes the intention of the user to write down the phrase and the intention of the user not to write down the phrase.

11. The method of claim 1, wherein the information items relate to traveling information and the set of intentions includes at least one of:

the intention to arrive at a predetermined time;

the intention to depart at a predetermined time; and

the intention to have predetermined minimum changing-over time.

12. A system for automatically responding to an inquiry from a user;

the system comprising;

a dialogue executor to execute a machine-controlled human-machine dialogue to determine a plurality of pre-determined query items specifying information to be verbally presented to the user;

a retriever to retrieve a plurality of information items from memory depending upon the query items;

a generator to generate at lest one natural language phrase to present the obtained information items according to a presentation scenario;

a verbal presentor to verbally present the generated phrase(s) to the user;

a determiner to determine a user'"'"'s intention from a predetermined set of intentions based upon utterance(s) of the user;

the intention reflecting a preferred way of presenting the information items; and

a selector to select the verbal presentation scenario from a predetermined set of presentation scenarios in dependence on the determined intention.

13. An article comprising a computer readable medium having instructions stored thereon which when executed causes:

execution of a machine-controlled human-machine dialogue to determine a plurality of predetermined query items specifying information to be verbally presented to the user;

retrieval of a plurality of information items from memory depending upon the query items;

determination of an intention of the user from a predetermined set of intentions, the intention reflecting a preferred way of presentation of the information items;

selection of the verbal presentation scenario from a predetermined set of presentation scenarios depending upon the determined intention, wherein the presentation scenarios are selected from the group consisting of semantic content of delivered information items, emphasis on or repetition of various informational items, and tailoring of actual vocabulary used;

generation of at least one natural language phrase to present the retrieved information items according to the presentation scenario; and

verbally presenting the generated phrase(s) to the user.