Computerized data-aware agent systems for retrieving data to serve a dialog between human user and computerized system

• US 10,387,536 B2
• Filed: 09/19/2012
• Issued: 08/20/2019
• Est. Priority Date: 09/19/2011
• Status: Active Grant

• Current Assignee: Personetics Technologies Limited (Personetics Technologies, Inc.)
• Original Assignee: Personetics Technologies Limited (Personetics Technologies, Inc.)
• Inventors: Govrin, David D., Sosna, David, Ophir, Ido, Vatnik, Dan
• Primary Examiner(s): Waldron, Scott A.
• Assistant Examiner(s): Lamardo, Viker A
• Application Number:
  US14/345,797
• Publication Number:
  US 20140229405A1
• Time in Patent Office: 2,526 Days
• Field of Search: 706 11
• CPC Class Codes: G06F 40/00 Handling natural language data speech analysis or synthesis, speech recognition G10L
• CPC Class Codes: G06N 5/02 Knowledge representation; Symbolic representation
• CPC Class Codes: G06N 5/04 Inference or reasoning models
• CPC Class Codes: G06Q 40/02 Banking, e.g. interest calculation or account maintenance credit or loans G06Q40/03
• CPC Class Codes: G10L 13/08 Text analysis or generation of parameters for speech synthesis out of text, e.g. grapheme to phoneme translation, prosody generation or stress or intonation determination
• CPC Class Codes: G10L 15/22 Procedures used during a speech recognition process, e.g. man-machine dialogue
• CPC Class Codes: G10L 2013/083 Special characters, e.g. punctuation marks
• CPC Class Codes: H04L 51/02 using automatic reactions or user delegation, e.g. automatic replies or chatbot-generated messages

Abstract

A system and method for data gathering system, comprising a data-aware knowledge base storing knowledge on relative costs of obtaining various data items; and a data retrieval decision-making processor operative, when an individual data element is sought to be retrieved, to determine whether or not to retrieve the data element by comparing at least one parameter representing need for the data element, also termed herein a utility value, with at least one parameter, retrieved from the data-aware knowledge base, which represents relative cost of obtaining the data element.

Claims

1. A data gathering system, the system comprising:

a virtual robot, including an automated chat system, operative to generate a dialog with a customer, via her or his mobile communication device, the robot being operative to employ external data from an enterprise systems to answer questions;

a data-aware knowledge base storing, in a machine-readable memory, knowledge on relative costs of obtaining various data items; and

a data retrieval decision-making processor operative, when an individual data element is sought to be retrieved by the virtual robot, to determine whether or not to retrieve said data element by comparing at least one parameter comprising a utility value representing a need for said data element with at least one parameter, retrieved from said data-aware knowledge base, which represents a relative cost of obtaining said data element,

wherein said data retrieval decision-making processor is configured to perform a reasoning process operative, given potential execution paths in a backward-chaining based expert system, to work from a set of desired goals, backward to dependent data items required to attain the goals, and

wherein said items include at least one of;

data retrieval from a computerized external information system, user dialog items which include posing direct questions to the user, and data processing only with no data retrieval,

wherein said utility value and a cost at each processing iteration are used to assess a cost-effectiveness of parameter links and a most cost-effective next item to resolve is found by optimization, and

wherein cost-effectiveness scores of the parameter links are reduced as an increasing function of the distance of the cost-effectiveness scores from a goal item to reflect said reasoning process'"'"'s backward-chaining nature,

thereby to provide a virtual robot configured for accessing the enterprise systems only when needed using a knowledge representation of data and reasoning process which compute the necessity and cost for retrieving a certain data element.

2. A system according to claim 1 wherein said knowledge base includes a hierarchy of costs wherein data which can be obtained without engaging either the user or an enterprise system serving the user is assigned a relatively low cost, data which is obtained by engaging the enterprise system is assigned a medium cost which exceeds said relatively low cost, and data which is obtained by engaging the user is assigned a high cost which exceeds said medium cost.

3. A system according to claim 1 which imitates human agent activity by accessing the systems only when needed using a knowledge representation of data and reasoning process.

4. A system according to claim 3 wherein said system computes the necessity and/or cost for retrieving at least one data element.

5. A system according to claim 3 wherein said system computes the cost for retrieving at least one data element.

6. A system according to claim 3 wherein the system takes into account that direct user questions and information gathering delays have an experiential impact on a human user'"'"'s experience, and selects an execution path accordingly.

7. A system according to claim 1 wherein said parameter representing need for said data element comprises a utility value.

8. A system according to claim 1 wherein the data retrieval decision-making processor is operative for computing cost-effectiveness scores for at least one parameter link.

9. A system according to claim 8 wherein negative cost-effectiveness scores are avoided and a Dijkstra-based search algorithm is used.

10. A system according to claim 8 wherein negative cost-effectiveness scores are avoided to enable spectral graph analysis.

11. A system according to claim 1 wherein said reasoning process is operative, at least once, given potential execution paths in a backward-chaining based expert system, to work from a set of desired goals, backward to dependent data items required to attain the goals, said items including data retrieval from a computerized external information system.

12. A system according to claim 1 wherein said reasoning process is operative, at least once, given potential execution paths in a backward-chaining based expert system, to work from a set of desired goals, backward to dependent data items required to attain the goals, said items including user dialog items which include posing direct questions to the user.

13. A data gathering method comprising:

reducing up-front data retrieval without embedding on-demand data collection into a reasoning process since due to large data sets and slow information systems, on-demand retrieval may have a negative effect on user experience as the user has to wait for new data to arrive, said reducing comprising;

storing, in a machine-readable memory, knowledge on relative costs of obtaining various data items, to generate a data-aware knowledge base; and

when an individual data element is sought to be retrieved, determining whether or not to retrieve said data element by comparing at least one parameter representing a need for said data element with at least one parameter, retrieved from said data-aware knowledge base, which represents a relative cost of obtaining said data element,

wherein said determining is performed by a data retrieval decision-making processor operative, given potential execution paths in a backward-chaining based expert system, to work from a set of desired goals, backward to identify dependent data items required to attain the goals, and

wherein said items include at least one of;

data retrieval from a computerized external information system, user dialog items which include posing direct questions to the user, and data processing only with no data retrieval,

wherein the method further comprises

identifying goals of the set of desired goals to serve as a basis for a shortest path tree search which returns a single-source shortest path for a graph with non-negative edge path costs, and

computing a current cost of identified items based on at least execution timing and determining a utility of at least one item parameter according to current data availability.

14. A method according to claim 13 and also comprising:

playing back human-generated queries previously executed in real-time and recorded, capturing additional information from said queries, and using said additional information to facilitate said determining.

15. A method according to claim 13 further comprising:

generating a topological representation of possible paths leading to at least one goal of the set of desired goals, wherein each path includes a sequence of nodes and each node is defined as including or not including data retrieval from an external source; and

using the topological representation to optimize dialog between a computerized system and a human user including selecting paths which reduce interaction with at least one computerized enterprise serving as an external source of data.

16. A method according to claim 13 further comprising:

generating a Laplacian matrix representation of a directed graph; and

examining the graph'"'"'s spectral decomposition including identifying a most cost-effective item which corresponds to a smallest non-zero matrix eigenvalue.

17. A method according to claim 13 wherein queries'"'"' execution durations are statistically analyzed to gradually learn average execution duration.

18. A method according to claim 13 wherein said data retrieval decision-making processor is operative, at least once, given potential execution paths in a backward-chaining based expert system, to work from a set of desired goals, backward to identify dependent data items required to attain the goals, said items including data retrieval from a computerized external information system.

19. A computer program product, comprising a non-transitory tangible computer-readable medium having computer-readable program code embodied therein, said computer-readable program code adapted to be executed to implement a data gathering method comprising:

storing in machine-readable memory, knowledge on relative costs of obtaining various data items, to generate a data-aware knowledge base; and

when an individual data element is sought to be retrieved, determining whether or not to retrieve said data element by comparing at least one parameter representing a need for said data element with at least one parameter, retrieved from said data-aware knowledge base, which represents a relative cost of obtaining said data element,

wherein said determining is performed by a data retrieval decision-making processor operative, given potential execution paths in a backward-chaining based expert system, to work from a set of desired goals, backward to identify dependent data items required to attain the goals, and

wherein said items include at least one of;

data retrieval from a computerized external information system, user dialog items which include posing direct questions to the user, and data processing only with no data retrieval,

wherein the method further comprises

computing a current cost of the identified items based on at least execution timing and determining a most cost-effective next item to resolve by optimization; and

locating and using a most cost-effective edge with a distance no larger than N item reference steps from a goal item as an approximation of optimal next item selection.

20. A computer program product according to claim 19 wherein said data retrieval decision-making processor is operative, at least once, given potential execution paths in a backward-chaining based expert system, to work from a set of desired goals, backward to identify dependent data items required to attain the goals, said items including data retrieval from a computerized external information system.