COLLABORATIVE LEARNING THROUGH USER GENERATED KNOWLEDGE

• US 20140201629A1
• Filed: 01/17/2013
• Published: 07/17/2014
• Est. Priority Date: 01/17/2013
• Status: Abandoned Application

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Heck, Larry
• Application Number:
  US13/743,511
• Publication Number:
  US 20140201629A1
• Time in Patent Office: Days
• US Class Current: 715/708
• CPC Class Codes: G06F 3/048 Interaction techniques based on graphical user interfaces [GUI]
• CPC Class Codes: G06N 5/00 Computing arrangements using knowledge-based models
• CPC Class Codes: G06N 5/022 Knowledge engineering; Knowledge acquisition

Abstract

A feedback loop is used by a central knowledge manager to obtain information from different users and deliver learned information to other users. Each user utilizes a personal assistant that learns from the user over time. The user may teach their personal assistant new knowledge through a natural user interface (NUI) and/or some other interface. For example, a combination of a natural language dialog and other non-verbal modalities of expressing intent (gestures, touch, gaze, images/videos, spoken prosody, . . . ) may be used to interact with the personal assistant. As knowledge is learned, each personal assistant sends the newly learned knowledge back to the knowledge manager. The knowledge obtained from the personal assistants is combined to form a collective intelligence. This collective intelligence is then transferred back to each of the individual personal assistants. In this way, the knowledge of one personal assistant benefits the other personal assistants through the feedback loop.

Claims

1. A method for collaborative learning using personal assistants, comprising:

receiving a user interaction at a personal assistant directed at performing a task;

determining when the personal assistant knows how to perform the task and when the personal assistant does not know how to perform the task, performing operations comprising;

learning instructions to perform the task using the personal assistant; and

sending the learned instructions to a knowledge manager that receives learned instructions from different personal assistants that are associated with different users and creates a collective user knowledge base comprising tasks that is shared with the personal assistants;

receiving information from the collective user knowledge base learned from interaction with personal assistants learned from other users.

2. The method of claim 1, wherein receiving the user interaction at the personal assistant directed at performing the task comprises receiving multimodal user input comprising speech input and at least one other form of input.

3. The method of claim 1, further comprising accessing a user-independent knowledge-base and extending the user-independent knowledge-base with the learned task.

4. The method of claim 1, wherein learning the instructions to perform the task using the personal assistant comprises creating a task model that is a graph that is constructed by mapping lower level concept sub graphs to higher level actions.

5. The method of claim 4, further comprising using at least one of:

a pattern recognition classifier, a sequential pattern recognition classifier; and

a hidden Markov model (HMM) to represent the task model

6. The method of claim 5, wherein the task model is initialized from search and browse logs comprising two or more of:

queries, clicks, page views, and dwell times.

7. The method of claim 3, further comprising determining a generalization of the task and extending the knowledge-base based on example data learned from the user.

8. The method of claim 1, wherein nodes of the knowledge-base are entities comprising:

a person, a place, and an item and edges of the knowledge-base are relations between the entities.

9. The method of claim 1, further comprising presenting a display to receive the instructions to perform the task from the user.

10. A computer-readable medium storing computer-executable instructions for collaborative learning using personal assistants, comprising:

receiving a user interaction at a personal assistant directed at performing a task;

determining when the personal assistant knows how to perform the task and when the personal assistant does not know how to perform the task, performing operations comprising;

learning instructions to perform the task using the personal assistant;

accessing a knowledge-base and extending the user-independent knowledge-base using the learned instructions to perform the task; and

sending the learned instructions to a knowledge manager that receives learned instructions from different personal assistants that are associated with different users and creates a collective user knowledge base comprising tasks that is shared with the personal assistants; and

receiving information from the collective user knowledge base learned from interaction with personal assistants learned from other users.

11. The computer-readable medium of claim 10, wherein receiving the user interaction at the personal assistant directed at performing the task comprises receiving multimodal user input comprising speech input and touch input.

12. The computer-readable medium of claim 10, wherein learning the instructions to perform the task using the personal assistant comprises creating a task model that is a graph that is constructed by mapping lower level concept sub graphs to higher level actions.

13. The computer-readable medium of claim 12, further comprising using a hidden Markov models (HMM) to represent the task model, where each state of the HMM is an intent.

14. The computer-readable medium of claim 13, wherein the task model is initialized from search and browse logs comprising two or more of:

queries, clicks, page views, and dwell times.

15. The computer-readable medium of claim 10, further comprising determining a generalization of the task and extending the knowledge-base based on example data learned from the user.

16. The computer-readable medium of claim 10, wherein nodes of the knowledge-base are entities comprising:

a person, a place, and an item and edges of the knowledge-base are relations between the entities.

17. A system for collaborative learning using personal assistants, comprising:

a processor and memory;

an operating environment executing using the processor;

a display; and

a knowledge manager that is configured to perform actions comprising;

receiving a user interaction at a personal assistant directed at performing a task;

determining when the personal assistant knows how to perform the task and when the personal assistant does not know how to perform the task, performing operations comprising;

learning instructions to perform the task using the personal assistant;

accessing a knowledge-base based on a determined generalization of the task and extending the user-independent knowledge-base using the learned instructions to perform the task; and

sending the learned instructions to a knowledge manager that receives learned instructions from different personal assistants that are associated with different users and creates a collective user knowledge base comprising tasks that is shared with the personal assistants; and

receiving information from the collective user knowledge base learned from interaction with personal assistants learned from other users.

18. The system of claim 17, wherein receiving the user interaction at the personal assistant directed at performing the task comprises receiving multimodal user input comprising speech input and touch input.

19. The system of claim 17, wherein learning the instructions to perform the task using the personal assistant comprises creating a task model that is a graph that is constructed by mapping lower level concept sub graphs to higher level actions.

20. The system of claim 19, further comprising initializing the task model from search and browse logs comprising two or more of:

queries, clicks, page views, and dwell times.