Internet expert system and method using free-form messaging in a dialogue format

• US 6,604,141 B1
• Filed: 10/12/1999
• Issued: 08/05/2003
• Est. Priority Date: 10/12/1999
• Status: Expired due to Term

• Current Assignee: NoHold, Inc.
• Original Assignee: NoHold, Inc.
• Inventors: Ventura, Diego
• Primary Examiner(s): An, Meng-Al T.
• Assistant Examiner(s): LIN, KENNY S
• Application Number:
  US09/416,663
• Time in Patent Office: 1,393 Days
• Field of Search: 709/202, 709/227, 709/220, 709/204, 707/4, 707/500, 707/3, 707/10, 707/6, 434/350, 434/118, 434/323, 434/362, 704/9, 704/10, 706/12, 706/53, 706/46, 706/54, 706/50, 706/16, 714/37
• US Class Current: 709/227
• CPC Class Codes: G06N 5/00 Computing arrangements using knowledge-based models
• CPC Class Codes: G06Q 10/10 Office automation; Time management
• CPC Class Codes: Y10S 707/99936 Pattern matching access

Abstract

Method of providing technical support using an expert system including exchanging information between the system and the computing devices in a conversational format utilizing the Internet. The expert system is connected to a knowledge base having data relevant to diagnosing operating conditions of at least one product. The knowledge base may have a tree-like arrangement of rules in which each rule includes a number of antecedents and a solution. Incoming text strings to the expert system are free-form descriptions of the operating conditions of the product for which technical support is sought. The outgoing text strings include questions strings and solutions strings, and the identification of user options, which may include requests for explanations and/or summaries of the session progress. Deduction and induction are employed to identify a rule which includes a solution that resolves the problem experienced by the user who is engaged in a session.

Claims

1. A method of providing technical support comprising steps of:

Storing a knowledge base in memory, including defining a tree arrangement of rules relating to said operating conditions, said rules including relationships between resolutions and antecedents of said resolutions, said knowledge base having data relevant to diagnosing operating conditions of at least one product;

Connecting said memory to an expert system for correlating resolutions to specific operating conditions and inputs that identify said specific operating conditions as applied to said at least one product;

Enabling Internet connectivity between said expert system and remote computing devices of users of said at least one product, including enabling a plurality of independent and concurrent sessions in which information is exchanged between said expert system and said remote computers in a conversational format, each said session being unrelated to the other said sessions concurrently being conducted and comprising steps of;

(a) transmitting outgoing text strings to a particular remote computer, said outgoing text strings being selected in an absence of human intervention and including question strings and resolution strings;

(b) receiving incoming text strings from said particular remote computer, said incoming text strings including free-form descriptions of operating conditions of said at least one product; and

(c) correlating said free-form descriptions within said incoming text strings and said data of said knowledge base including identifying indications of said antecedents contained within said incoming text strings from said particular remote computer, wherein during execution of said correlating, said step of transmitting resolution strings include basing content of said resolution strings upon identification of correlations

wherein said steps of transmitting and correlating enable forward deducing from an indication of a first antecedent to a related first resolution and enable backward inducing in which an outgoing question string is transmitted to request confirmation of presence of a second antecedent that is also related to said first resolution, said second antecedent having been previously unidentified in any said incoming text string.

2. The method of claim 1 wherein said step of transmitting outgoing text strings includes transmitting explanation strings that are each responsive to an explanation request from said particular remote computer and that are each indicative of a rationale for selecting and transmitting a question string that was previously transmitted to said particular remote computer.

3. The method of claim 1 further comprising steps of maintaining a log of each said session and generating reports regarding said sessions, such that said reports may be utilized to enhance performance of said expert system.

4. The method of claim 1 further comprising a step of transmitting a summary to each said remote computer from which a summary request is received during one of said sessions.

5. The method of claim 1 wherein said step of enabling Internet connectivity includes providing browser-based access via the World Wide Web (WWW).

6. The method of claim 1 wherein said step of storing said knowledge base includes providing data related to operating conditions for a specific service, said specific service being said product.

7. A method of providing technical support comprising steps of:

forming a knowledge base in memory, including defining a tree arrangement of rules in which each rule includes a solution and at least one related antecedent that represents a condition which indicates appropriateness of said solution;

enabling access to said memory via the World Wide Web, including enabling computer-executed exchanges that include;

(a) receiving free-form messages having descriptive content indicative of conditions of a product of interest;

(b) electronically reading said free-form messages;

(c) determining whether said free-form messages describe conditions that are representative of one of said antecedents within said tree arrangement;

(d) if a match is identified between a specific antecedent and description of conditions within a particular free-form message, transmitting a signal that presents a grammatically correct solution message at a source of said particular free-form message, said solution message being formed to propose the solution to which said specific antecedent is related;

(e) if a match is not identified between an antecedent and said description of conditions, transmitting a signal that presents a grammatically correct question message at said source, including selecting said question message based upon inducing additional description in a responsive free-form message from said source; and

(f) repeating steps (a)-(e) for said responsive free-form message;

wherein said steps (a)-(f) are implemented to simulate a dialogue and are executed in an absence of human intervention.

8. The method of claim 7 wherein said step of selecting said question message includes selecting a question content and selecting from a plurality of available options for phrasing said question content, said selecting from said available options being implemented to provide variety in presentations of said question messages.

9. The method of claim 7, further comprising steps of logging each session with sources of said free-form messages and enabling a particular session to subsequently resume following a termination of said steps (a)-(f).

10. The method of claim 7 wherein said steps (a)-(f) are implemented such that said exchanges are in a format compatible with communications with WWW browser applications at sources of said free-form messages.

11. A system for providing product support by exchanges of messages that simulate free-form dialogue comprising:

memory having a stored knowledge base that defines a tree arrangement of rules, said rules having solutions and having antecedents which are related to said solutions;

means for communicating incoming and outgoing text strings in a format compatible with transmissions via the World Wide Web (WWW);

reading means for receiving said incoming text strings via said communicating means and for recognizing contents of said incoming text strings, said reading means having a sufficiently large search vocabulary to enable said incoming text strings to be free-form messages;

deduction means connected to said reading means and said memory for detecting correlations between said contents of said incoming text strings and said rules; and

message means for transferring said outgoing text strings to said communicating means in response to operations of said deduction means, said outgoing text strings including solution messages that identify solutions when correlations are detected between said contents of said incoming text strings and said rules, said message means being enabled to transfer a specific induction message to said communicating means in response to a determination at said deduction means that confirmation of a correlation is required, said specific induction message being representative of a question having content which is intended to acquire said confirmation.

12. The system of claim 11 further comprising a means for logging sessions of said incoming and outgoing text strings with specific remote computers.

13. The system of claim 12 wherein said deduction means and said logging means are cooperative to enable resumption of a specific session subsequent to a first termination of said specific session.

14. The system of claim 11 wherein said deduction means and said message means are cooperative to trigger one of said specific induction messages requesting confirmation of a second antecedent to a particular rule when a preceding incoming text string is read to confirm a first antecedent of said particular rule.

15. The system of claim 11 wherein said reading means, deduction means and message means are implemented in computer software that includes a read module, a deduction module and an induction module.