SPECTRAL DIAGNOSTIC ENGINE FOR CUSTOMER SUPPORT CALL CENTER

• US 20160173690A1
• Filed: 12/12/2014
• Published: 06/16/2016
• Est. Priority Date: 12/12/2014
• Status: Active Grant

• Current Assignee: Conduent Business Services, LLC (Conduent, Inc.)
• Original Assignee: Xerox Corporation (Xerox Holdings Corp.)
• Inventors: Perez, Julien, Bouchard, Guillaume
• Granted Patent: US 9,813,555 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G06F 11/0748 in a remote unit communicating with a single-box computer node experiencing an error/fault remote testing G06F11/2294
• CPC Class Codes: G06F 11/079 Root cause analysis, i.e. error or fault diagnosis in a hardware test environment G06F11/22; in a software test environment G06F11/36
• CPC Class Codes: G06F 17/16 Matrix or vector computation , e.g. matrix-matrix or matrix-vector multiplication, matrix factorization matrix transposition G06F7/78
• CPC Class Codes: G06Q 10/20 Administration of product repair or maintenance
• CPC Class Codes: G06Q 30/01 Customer relationship services
• CPC Class Codes: H04M 3/5175 Call or contact centers supervision arrangements

Abstract

A collective matrix is constructed, having a diagnostic sessions dimension and a diagnostic state descriptors dimension. The diagnostic state descriptors dimension includes a plurality of symptom fields, a plurality of root cause fields, and a plurality of solution fields. Collective matrix factorization of the collective matrix is performed to generate a factored collective matrix comprising a sessions factor matrix embedding diagnostic sessions and a descriptors factor matrix embedding diagnostic state descriptors. An in-progress diagnostic session is embedded in the factored collective matrix. A symptom or solution is recommended for evaluation in the in-progress diagnostic session based on the embedding. The diagnostic state descriptors dimension may further include at least one information field storing a representation (for example, a bag-of-words representation) of a semantic description of a problem being diagnosed by the in-progress diagnostic session.

Claims

1. A diagnostic engine comprising:

an electronic data processing device programmed to perform a diagnostic support method including the operations of;

constructing a collective matrix having a diagnostic sessions dimension and a diagnostic state descriptors dimension wherein the diagnostic state descriptors dimension includes at least a plurality of symptom fields corresponding to the symptoms of a symptoms set and a plurality of solution fields corresponding to the solutions of a solutions set;

performing collective matrix factorization of the collective matrix to generate a factored collective matrix comprising a sessions factor matrix embedding diagnostic sessions and a descriptors factor matrix embedding diagnostic state descriptors;

embedding an in-progress diagnostic session in the factored collective matrix; and

determining one of a predicted solution for the in-progress diagnostic session and a not-yet-acquired probative symptom for the in-progress diagnostic session based on the embedding.

2. The diagnostic engine of claim 1 wherein the wherein the diagnostic state descriptors dimension further includes a plurality of root cause fields corresponding to the root causes of a root causes set.

3. The diagnostic engine of claim 1 wherein the diagnostic state descriptors dimension further includes at least one information field storing information about a problem being diagnosed by the in-progress diagnostic session.

4. The diagnostic engine of claim 3 wherein the at least one diagnostic session information field includes one or more fields storing a bag-of-words representation of a semantic description of the problem being diagnosed by the in-progress diagnostic session.

5. The diagnostic engine of claim 1 wherein the diagnostic support method includes the further operation of:

in response to determining a predicted solution for the in-progress diagnostic session, displaying the predicted solution for the in-progress diagnostic session on a display device.

6. The diagnostic engine of claim 1 wherein the diagnostic support method includes the further operation of:

in response to determining a not-yet-acquired probative symptom, displaying a recommendation to acquire the value of the probative symptom for the in-progress diagnostic session on a display device.

7. The diagnostic engine of claim 1 wherein the operation of performing collective matrix factorization comprises:

evaluating

8. The diagnostic engine of claim 7 wherein the evaluating comprises:

performing an alternating least squares optimization in which the following are computed alternately;

A*=arg min_A∥

(Y−

AB)W∥

+λ

_a∥

A∥

and

B*=arg min_B∥

(Y−

AB)W∥

+λ

_b∥

B∥

.

9. The diagnostic engine of claim 1 wherein the determining operation comprises determining a predicted solution for the in-progress diagnostic session by:

determining the solution of the solutions set that maximizes information gain for the embedding.

10. The diagnostic engine of claim 1 wherein the determining operation comprises determining a not-yet-acquired probative symptom for the in-progress diagnostic session by:

determining the not-yet-acquired symptom of the symptoms set that maximizes information gain for the embedding.

11. A diagnostic call center comprising:

a diagnostic engine as set forth in claim 1; and

a call center device providing two-way communication between a customer and a call center agent to conduct the in-progress diagnostic session, the call center device further providing a user interface configured to display to the call center agent the predicted solution or not-yet-acquired probative symptom determined by the diagnostic engine for the in-progress diagnostic session being conducted via the call center device.

12. A diagnostic method comprising:

constructing a collective matrix Y having a diagnostic sessions dimension and a diagnostic state descriptors dimension wherein the diagnostic state descriptors dimension includes at least a plurality of symptom fields corresponding to the symptoms of a symptoms set S and a plurality of solution fields corresponding to the solutions of a solutions set ;

performing collective matrix factorization of the collective matrix to generate a factored collective matrix comprising a sessions factor matrix A embedding diagnostic sessions and a descriptors factor matrix B embedding diagnostic state descriptors;

embedding an in-progress diagnostic session in the factored collective matrix; and

generating a recommended symptom or solution to evaluate for the in-progress diagnostic session based on the embedding;

wherein the constructing, performing, embedding, and generating operations are performed by an electronic data processing device.

13. The diagnostic method of claim 12 wherein the wherein the diagnostic state descriptors dimension further includes a plurality of root cause fields corresponding to the root causes of a root causes set C.

14. The diagnostic method of claim 12 wherein the diagnostic state descriptors dimension further includes at least one information field storing a representation of a semantic description of a problem being diagnosed by the in-progress diagnostic session.

15. The diagnostic method of claim 12 wherein the operation of performing collective matrix factorization comprises:

evaluating

16. The diagnostic method of claim 12 wherein the evaluating comprises:

performing an alternating least squares optimization in which the following are computed alternately;

17. The diagnostic method of claim 12 further comprising:

displaying the recommended symptom or solution to evaluate on a call center device to be viewed by a call center staffer; and

conducting the in-progress diagnostic session using the call center device for two-way communication between the call center staffer and a customer, the two-way communication including communicating the recommended symptom or solution from the call center staffer to the customer.

18. The diagnostic method of claim 12 wherein the generating comprises:

generating a recommended symptom as the symptom of the set of symptoms S that maximizes information gain of the embedding.

19. The diagnostic method of claim 12 wherein the generating comprises:

generating a recommended solution as the solution of the set of solutions L that maximizes information gain of the embedding.

20. A non-transitory storage medium storing instructions readable and executable by a computer to perform a diagnostic method comprising:

performing collective matrix factorization of a collective matrix having a diagnostic sessions dimension and a diagnostic state descriptors dimension wherein the diagnostic state descriptors dimension includes at least a plurality of symptom fields corresponding to the symptoms of a symptoms set , a plurality of root cause fields corresponding to the root causes of a root causes set , and a plurality of solution fields corresponding to the solutions of a solutions set ; and

generating a recommended symptom or solution to evaluate for an in-progress diagnostic session using the factored collective matrix.