Granted Patent B2

US 12,705,162 · App. 18/522,508 · Granted Aug 11, 2026

Inventors: Desmond Fitzpatrick (Ossining, NY); Andrew C. M. Hicks (Highland, NY); Michael Terrence Cohoon (Fishkill, NY)

Assignee: International Business Machines Corporation

G06F11/3608

Abstract

Claims (52)

1 . A method comprising:

identifying, during execution of an application including a portion of multipathing code comprising a first path and a second path, a first fault in an execution of the first path,

wherein the first fault includes one or more execution errors resulting from the execution of the first path;

generating, based on the first fault, a first prompt for an artificial intelligence (AI) language model,

wherein generating the first prompt for the AI language model includes:

generating the first prompt based on whether the first fault resulted from AI-generated code or human-generated code,

wherein, when the first fault resulted from human-generated code, the first prompt is generated to include a request to generate new code based on the human-generated code; and

applying a first modification to the first path based on an output of the AI language model in response to the first prompt.

2 . The method of claim 1 , wherein the second path is executed in response to identifying the first fault in the first path; and wherein generating the first prompt is further based on determining that the second path executed successfully.

3 . The method of claim 1 , wherein the first prompt identifies a failure point in the first path.

4 . The method of claim 1 , wherein the first path is a default code path and the second path is a secondary code path.

5 . The method of claim 1 , wherein first source code corresponding to the first path is based on a first programming language and second source code corresponding to the second path is based on a second programming language.

6 . The method of claim 1 , wherein the first path and the second path are directed to identical outcomes.

7 . The method of claim 1 , wherein, when the first fault resulted from AI-generated code, the first prompt is generated to include a request to regenerate the AI-generated code based on failure state information associated with the AI-generated code.

8 . The method of claim 1 further comprising:

identifying a second fault during execution of the first path after application of the first modification, wherein the second fault resulted from a same portion of code as the first fault;

generating a second prompt for the AI language model based on failure information of the first fault and the second fault; and

applying a second modification to the first path based on the output of the AI language model in response to the second prompt.

9 . The method of claim 1 further comprising:

identifying a second fault during execution of the first path after application of the first modification;

adjusting one or more parameters of the AI language model; and

generating a second prompt for the AI language model.

10 . The method of claim 1 further comprising:

in response to determining that execution of the first path after application of the first modification progressed beyond a point of the first fault, associating the first prompt with a resolution of the first fault.

11 . An apparatus comprising:

a processing device; and

memory operatively coupled to the processing device, wherein the memory stores computer program instructions that, when executed, cause the processing device to:

identify, during execution of an application including a portion of multipathing code comprising a first path and a second path, a first fault in an execution of the first path,

wherein the first fault is identified during the execution of the first path, and

wherein the first fault includes one or more execution errors resulting from the execution of the first path;

generate, based on the first fault, a first prompt for an artificial intelligence (AI) language model,

wherein the computer program instructions that cause the processing device to generate the first prompt for the AI language model further cause the processing device to:

generate the first prompt based on whether the first fault resulted from AI-generated code or human-generated code,

 wherein, when the first fault resulted from human-generated code, the first prompt is generated to include a request to generate new code based on the human-generated code; and

apply a first modification to the first path based on an output of the AI language model in response to the first prompt.

12 . The apparatus of claim 11 , wherein the second path is executed in response to identifying the first fault in the first path; and wherein the first prompt is generated further based on determining that the second path executed successfully.

13 . The apparatus of claim 11 , wherein, when the first fault resulted from AI-generated code, the first prompt is generated to include a request to regenerate the AI-generated code based on failure state information associated with the AI-generated code.

14 . The apparatus of claim 11 further comprising instructions that, when executed, cause the processing device to:

identify a second fault during execution of the first path after application of the first modification, wherein the second fault resulted from a same portion of code as the first fault;

generate a second prompt for the AI language model based on failure information of the first fault and the second fault; and

apply a second modification to the first path based on the output of the AI language model in response to the second prompt.

15 . The apparatus of claim 11 further comprising instructions that, when executed, cause the processing device to:

identify a second fault during execution of the first path after application of the first modification;

adjust one or more parameters of the AI language model; and

generate a second prompt for the AI language model.

16 . A computer program product comprising a computer readable storage medium, wherein the computer readable storage medium comprises a computer program that, when executed:

identifies, during execution of an application including a portion of multipathing code comprising a first path and a second path, a first fault in an execution of the first path;

generates, based on the first fault, a first prompt for an artificial intelligence (AI) language model,

wherein the computer program that, when executed to generate the first prompt for the AI language model, further:

generates the first prompt based on whether the first fault resulted from AI-generated code or human-generated code,

wherein, when the first fault resulted from human-generated code, the first prompt is generated to include a request to generate new code based on the human-generated code; and

applies a first modification to the first path based on an output of the AI language model in response to the first prompt.

Assignments (1)

Continuity (1)

References Cited (15)