Crafting feedback dialogue with a digital assistant

• US 10,446,142 B2
• Filed: 05/20/2015
• Issued: 10/15/2019
• Est. Priority Date: 05/20/2015
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Inventors: Lim, Melissa Nicole, Mitchell, Margaret Ann, Quirk, Christopher Brian
• Primary Examiner(s): Chuang, Jung-Mu T
• Application Number:
  US14/718,071
• Publication Number:
  US 20160342317A1
• Time in Patent Office: 1,609 Days
• CPC Class Codes: G06F 40/30 Semantic analysis
• CPC Class Codes: G06Q 10/00 Administration; Management
• CPC Class Codes: G10L 15/1822 Parsing for meaning understanding
• CPC Class Codes: G10L 15/22 Procedures used during a speech recognition process, e.g. man-machine dialogue
• CPC Class Codes: G10L 2015/0635 updating or merging of old and new templates; Mean values; Weighting
• CPC Class Codes: G10L 2015/0638 Interactive procedures

Abstract

Examples described herein dynamically personalize a digital assistant for a specific user, creating a personal connection between the digital assistant and the user. The digital assistant accesses user activity and generates queries based on the user activity. The digital assistant facilitates natural language conversations as machine learning sessions between the digital assistant and the user using the one or more queries to learn the user'"'"'s preferences and receives user input from the user during the learning session in response to the queries. The digital assistant dynamically updates a personalized profile for the user based on the user input during the natural language conversations.

Claims

1. A system for personalizing a digital assistant, the system comprising:

a processing unit;

a memory communicatively coupled to the processing unit, the memory maintaining activity logs corresponding to one or more computing device components of one or more user devices; and

the digital assistant, executed by the processing unit, the digital assistant further executed to;

analyze at least one user activity of logged user activities from the activity logs maintained in the memory, the logged user activities based on signals generated via the one or more computing device components of the one or more user devices;

generate one or more inferences related to user preferences based on the analyzed at least one user activity of logged user activities from the activity logs;

generate at least one question directed towards confirming at least one generated inference of the one or more generated inferences;

store the at least one generated question in a learning queue;

determine whether the learning queue has reached a threshold number of stored questions;

responsive to determining that the learning queue has reached the threshold number of stored questions, trigger a learning session associated with personalizing the digital assistant by presenting one or more questions from the learning queue via a natural language conversation;

receive a user input during the learning session responsive to the one or more questions;

identify a sentiment associated with at least one facet of the received user input; and

update a user profile based on the received user input and the identified sentiment of the received user input to confirm or deny the at least one generated inference.

2. The system of claim 1, wherein the one or more computing device components include at least one of a global positioning system, an Internet browser, a sentiment classifier, an electronic messaging client, a text messaging client, a camera, an error log, or a personalized knowledge base.

3. The system of claim 1 wherein the digital assistant is further executed to:

generate a response during the learning session based at least in part on the identified sentiment.

4. The system of claim 1 wherein the digital assistant is further executed to:

label the at least one facet of the received user input according to the identified sentiment; and

update the user profile with the labeled facet.

5. The system of claim 1, wherein the threshold is a number of areas of interest having at least one corresponding generated question.

6. The system of claim 1, wherein the trigger is further based on detected user inactivity.

7. The system of claim 1, wherein the trigger is further based on time identified in a user'"'"'s schedule as free time.

8. A mobile device comprising:

a processing unit;

a memory communicatively coupled to the processing unit; and

a digital assistant executed by the processing unit to;

process activity logs maintained on the memory to identify one or more areas of desired feedback associated with one or more inferences developed at the digital assistant based on the activity logs, the activity logs being configured to store at least one activity of a user based on signals generated via one or more computing device components of the mobile device;

generate one or more questions directed towards confirming the one or more inferences developed at the digital assistant corresponding to the one or more areas of desired feedback;

store the generated one or more questions in a learning queue;

determine whether the learning queue has reached a threshold number of stored questions;

responsive to determining that the learning queue has reached the threshold number of stored questions, trigger a learning session associated with the one or more inferences developed at the digital assistant via a natural language conversation based on the one or more generated questions stored in the learning queue;

receive user input during the natural language conversation responsive to the one or more questions;

identify a sentiment associated with at least one facet of the received user input;

confirm or deny at least one inference of the one or more inferences developed at the digital assistant based on the received user input and the identified sentiment; and

perform at least one task related to a user preference based on the at least one confirmed or denied inference.

9. The mobile device of claim 8 wherein the digital assistant is further executed to:

generate one or more other inferences related to user preferences based on the activity logs; and

generate at least one question directed towards a task associated with at least one of the one or more other inferences.

10. The mobile device of claim 8 further comprising:

a global positioning system generating one or more signals corresponding to at least one location associated with the mobile device, the one or more signals stored in the activity logs; and

the digital assistant further executed to process the one or more signals to identify the one or more areas of desired feedback.

11. The mobile device of claim 8 further comprising:

a browser executed by the processing unit, the browser generating one or more signals corresponding to browser history associated with the mobile device, the one or more signals stored in the activity logs; and

the digital assistant further executed to process the one or more signals to identify the one or more areas of desired feedback.

12. The mobile device of claim 8 further comprising:

an electronic messaging client executed by the processing unit, the electronic messaging client generating one or more signals corresponding to a text input associated with the mobile device, the one or more signals stored in the activity logs; and

the digital assistant further executed to process the one or more signals to identify the one or more areas of desired feedback.

13. The mobile device of claim 8 further comprising:

a camera generating one or more signals corresponding to at least one captured image, the one or more signals stored in the activity logs; and

the digital assistant further executed to process the one or more signals to identify the one or more areas of desired feedback.

14. The mobile device of claim 8 further comprising:

an error log generating one or more signals corresponding to at least one user-initiated task failure, the one or more signals stored in the activity logs; and

the digital assistant further executed to process the one or more signals to identify the one or more areas of desired feedback.

15. A method for digital assistant personalization, the method comprising:

analyzing at least one user activity from logged user activities of activity logs maintained in a memory of a computing device, the logged user activities based on signals generated via one or more computing device components of the computing device;

generating one or more inferences related to user preferences based on the analyzed at least one user activity of logged user activities from the activity logs;

generating at least one question directed towards confirming at least one generated inference of the generated one or more inferences to personalize the digital assistant to a user of the computing device based on the one or more generated inferences from the activity logs;

storing the generated at least one question in a learning queue;

determining whether the learning queue has reached a threshold number of stored questions;

responsive to determining that the learning queue has reached the threshold number of stored questions, triggering, by the digital assistant, a machine learning session via a natural language conversation between the digital assistant and the user of the computing device using the number of stored questions in the learning queue;

receiving a user input during the natural language conversation in response to the number of stored questions;

identifying sentiment associated with at least one facet of the received user input; and

updating a personalized profile associated with the user based on the received user input and the identified sentiment to confirm or deny the at least one generated inference in order to personalize the digital assistant.

16. The method of claim 15 further comprising:

generating a response to the user input during the natural language conversation of the natural language conversation based at least in part on the identified sentiment.

17. The method of claim 15 wherein generating the at least one question further comprises:

processing the activity logs, including performing a keyword recognition on at least one of browser history data, text messaging data, electronic messaging data, or social media application data.

18. The method of claim 15 wherein generating the at least one question further comprises:

processing the activity logs, including identifying one or more locations corresponding to a computing device location history using a global positioning system data associated with the computing device and a search engine.

19. The method of claim 15 wherein generating the at least one question further comprises:

processing the activity logs, including detecting one or more images captured by a camera associated with the computing device and identifying at least one entity in at least one of the one or more images.

20. The method of claim 15 wherein generating the at least one question further comprises:

processing the activity logs, including identifying at least one failed user-initiated task.