System and method providing utility-based decision making about clarification dialog given communicative uncertainty

• US 7,580,908 B1
• Filed: 04/07/2005
• Issued: 08/25/2009
• Est. Priority Date: 10/16/2000
• Status: Expired due to Fees

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Paek, Timothy S., Horvitz, Eric J.
• Primary Examiner(s): Holmes; Michael B
• Application Number:
  US11/100,699
• Time in Patent Office: 1,601 Days
• Field of Search: 706/45
• US Class Current: 706/45
• CPC Class Codes: G06N 7/01 Probabilistic graphical models, e.g. probabilistic networks
• CPC Class Codes: G10L 15/1822 Parsing for meaning understanding

Abstract

An interactive communications system is provided in accordance with the present invention. The system includes a component for analyzing sequential communications including speech, gestures and other modalities related to an underlying communicative intention. The component concurrently employs at least two of the communications in determining an action to facilitate achieving the intention.

Claims

1. An interactive command system, comprising the following:

a processor;

a memory communicatively coupled to the processor, the memory having stored therein computer-executable instructions to implement the system including;

a component to assess reliability of a human communication;

a component to infer a probability associated with an intent of the communication; and

a component that performs an action with a highest utility determined by employing a utility based probabilistic decision model based upon the inferred probability to facilitate achieving the intent of the communication.

2. The system of claim 1, wherein the action is a clarification action.

3. The system of claim 2, wherein the clarification action includes at least one of asking user to repeat utterance, note a recognition of a users hesitation and try again, note a reception of a noise and inquire, getting the users intention, apologizing for an interruption and forgoing the action, engage troubleshooting regarding the clarification action.

4. The system of claim 1, wherein the actions are determined from a confidence threshold associated with an expected utility.

5. The system of claim 1, wherein the action is a domain level action.

6. The system of claim 5, wherein the domain-level action includes at least one of engage, disengage, enable and disable.

7. The system of claim 1, wherein the intent of the communication is associated with at least one of acknowledgement, negation, reflection, unrecognized response, and no response.

8. The system of claim 7, wherein a plurality of candidate utterances are defined for each intention.

9. The system of claim 7, wherein the intent of the communications is influenced by the reliability of the communications.

10. The system of claim 9, wherein the reliability of communications are at least one of related to noise and background input levels associated with the communications.

11. The system of claim 8, wherein the plurality of candidate utterances are analyzed according to a Bayesian network model, the model is determined over different instances of time to probabilistically infer a communicators intent.

12. The system of claim 1, further comprising an animated agent to interact with a communicator to determine the communicator'"'"'s intent.

13. The system of claim 12, further comprising the animated agent provides services relating to synchronization of contacts, appoints, and tasks, associated with at least one of an e-mail, scheduling, calendaring, and planning service.

14. The system of claim 1, further comprising a development tool that is utilized for a plurality of command and control domains.

15. A method for interactive human communications, comprising the following:

employing a processor executing computer executable instructions to perform the following acts;

assessing reliability of a human communication;

inferring probability associated with an intent of the communication; and

performing an action with a highest utility determined by employing a probabilistic decision model based upon the inferred probability to facilitate achieving the communicative intention.

16. An interactive system, comprising the following components:

a processor;

a memory communicatively coupled to the processor, the memory having stored therein computer-executable instructions to implement the system including;

a first component that analyzes a sequential human communications related to an underlying communicative intention, the component concurrently employing a utility based Bayesian decision model utilizing at least two of the communications to determine an action with a highest utility to facilitate achieving the intention; and

a second component that performs the action when the underlying communicative intention is above a confidence threshold.

17. The system of claim 16, wherein the first component includes speech, gestures and other modalities related to an underlying communicative intention, the component concurrently employing at least two of the communications modalities in determining an action to facilitate achieving the intention.