Knowledge graph weighting during chatbot sessions

Patent metadata

Field	Value
Publication number	US-11948099-B2
Application number	US-202117460674-A
Country	US
Kind code	B2
Filing date	Aug 30, 2021
Priority date	Aug 30, 2018
Publication date	Apr 2, 2024
Grant date	Apr 2, 2024

Abstract

Implementations include providing, by the PKG platform, an initial knowledge graph based on user-specific data associated with a user, and a domain-specific knowledge graph, receiving, by the PKG platform, data representative of at least one answer provided from the user to a respective question, providing, by the PKG platform, an expanded knowledge graph based on the initial knowledge graph, the expanded knowledge graph including one or more nodes and respective edges based on the data, generating, by the PKG platform, a weighted knowledge graph based a groundtruth knowledge graph, and a targeted knowledge graph, the groundtruth knowledge graph including one or more true answers, and the targeted knowledge graph including the at least one answer provided from the user, and generating, by the PKG platform, the hyper-personalized knowledge graph (hpKG) based on the weighted knowledge graph, the hpKG being unique to the user within a domain.

First claim

What is claimed is:

1. A computer-implemented method executed by one or more processors for providing a hyper-personalized knowledge graph using a personal knowledge graph (PKG) platform, the method comprising:
providing an expanded knowledge graph based on an initial knowledge graph, the expanded knowledge graph comprising one or more nodes and respective edges based on data representative of at least one answer provided from a user to a respective question;
generating a weighted knowledge graph based a groundtruth knowledge graph, and a targeted knowledge graph, the groundtruth knowledge graph comprising one or more true answers, and the targeted knowledge graph comprising the at least one answer provided from the user; and
generating the hyper-personalized knowledge graph (hpKG) based on the weighted knowledge graph, the hpKG being unique to the user within a domain.

2. The method of claim 1 , further comprising identifying information in a domain-specific knowledge graph, and adding a node to the initial knowledge graph based on the identified information to provide the expanded knowledge graph.

3. The method of claim 1 , wherein the groundtruth knowledge graph is generated based on a domain-specific knowledge graph.

4. The method of claim 1 , wherein the groundtruth knowledge graph is generated by adding one or more nodes to a domain-specific knowledge graph.

5. The method of claim 1 , wherein generating the weighted knowledge graph comprises:
converting the groundtruth knowledge graph to a multi-dimensional vector by graph embedding to provide one or more dimensional values for each node of the groundtruth knowledge graph; and
applying at least a portion of the one or more dimensional values to the targeted knowledge graph to provide one or more weights for respective edges.

6. The method of claim 1 , wherein a weight of the weighted knowledge graph is determined based on a correctness of the at least one answer determined from the groundtruth knowledge graph.

7. The method of claim 1 , wherein the weighted knowledge graph comprises a node and an edge that are included in the groundtruth knowledge graph, and that are absent from the targeted knowledge graph.

8. The method of claim 7 , wherein the node represents a concept, and a question relating to the concept is transmitted to the user in a subsequent iteration.

9. The method of claim 1 , wherein the hpKG is provided by integrating the weighted knowledge graph, and a context-based knowledge graph.

10. The method of claim 1 , further comprising iteratively questioning the user based on a set of questions comprising the respective question to provide a set of responses comprising the at least one response.

11. The method of claim 1 , wherein the data is received through one or more chatbot-based interactions with the user.

12. A non-transitory computer-readable storage medium coupled to one or more processors and having instructions stored thereon which, when executed by the one or more processors, cause the one or more processors to perform operations for providing a hyper-personalized knowledge graph using a personal knowledge graph (PKG) platform, operations comprising:
providing an expanded knowledge graph based on an initial knowledge graph, the expanded knowledge graph comprising one or more nodes and respective edges based on data representative of at least one answer provided from a user to a respective question;
generating a weighted knowledge graph based a groundtruth knowledge graph, and a targeted knowledge graph, the groundtruth knowledge graph comprising one or more true answers, and the targeted knowledge graph comprising the at least one answer provided from the user; and

generating the hyper-personalized knowledge graph (hpKG) based on the weighted knowledge graph, the hpKG being unique to the user within a domain.

13. The non-transitory computer-readable storage medium of claim 12 , wherein operations further comprise identifying information in a domain-specific knowledge graph, and adding a node to the initial knowledge graph based on the identified information to provide the expanded knowledge graph.

14. The non-transitory computer-readable storage medium of claim 12 , wherein the groundtruth knowledge graph is generated based on a domain-specific knowledge graph.

15. The non-transitory computer-readable storage medium of claim 12 , wherein the groundtruth knowledge graph is generated by adding one or more nodes to a domain-specific knowledge graph.

16. The non-transitory computer-readable storage medium of claim 12 , wherein generating the weighted knowledge graph comprises:
converting the groundtruth knowledge graph to a multi-dimensional vector by graph embedding to provide one or more dimensional values for each node of the groundtruth knowledge graph; and
applying at least a portion of the one or more dimensional values to the targeted knowledge graph to provide one or more weights for respective edges.

17. The non-transitory computer-readable storage medium of claim 12 , wherein a weight of the weighted knowledge graph is determined based on a correctness of the at least one answer determined from the groundtruth knowledge graph.

18. The non-transitory computer-readable storage medium of claim 12 , wherein the weighted knowledge graph comprises a node and an edge that are included in the groundtruth knowledge graph, and that are absent from the targeted knowledge graph.

19. The non-transitory computer-readable storage medium of claim 18 , wherein the node represents a concept, and a question relating to the concept is transmitted to the user in a subsequent iteration.

20. The non-transitory computer-readable storage medium of claim 12 , wherein the hpKG is provided by integrating the weighted knowledge graph, and a context-based knowledge graph.

21. The non-transitory computer-readable storage medium of claim 12 , wherein operations further comprise iteratively questioning the user based on a set of questions comprising the respective question to provide a set of responses comprising the at least one response.

22. The non-transitory computer-readable storage medium of claim 12 , wherein the data is received through one or more chatbot-based interactions with the user.

23. A system, comprising:
one or more processors; and
a computer-readable storage device coupled to the one or more processors and having instructions stored thereon which, when executed by the one or more processors, cause the one or more processors to perform operations for providing a hyper-personalized knowledge graph using a personal knowledge graph (PKG) platform, operations comprising:
providing an expanded knowledge graph based on an initial knowledge graph, the expanded knowledge graph comprising one or more nodes and respective edges based on data representative of at least one answer provided from a user to a respective question;
generating a weighted knowledge graph based a groundtruth knowledge graph, and a targeted knowledge graph, the groundtruth knowledge graph comprising one or more true answers, and the targeted knowledge graph comprising the at least one answer provided from the user; and
generating the hyper-personalized knowledge graph (hpKG) based on the weighted knowledge graph, the hpKG being unique to the user within a domain.

24. The system of claim 23 , wherein operations further comprise identifying information in a domain-specific knowledge graph,

Assignees

• Accenture Global Solutions Ltd

Inventors

• Lecue Freddy
• Gueret Christophe Dominique Marie
• Cahalane Diarmuid John

Classifications

• G06N5/04 Primary

  Inference or reasoning models · CPC title

• G06F16/9024

  Graphs; Linked lists (G06F16/9027 takes precedence) · CPC title

• G06N3/006

  based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO] · CPC title

• G06F16/90332 Primary

  Natural language query formulation or dialogue systems · CPC title

• H04L51/02

  using automatic reactions or user delegation, e.g. automatic replies or chatbot-generated messages · CPC title