Granted Patent B2

US 12,669,991 · App. 18/508,381 · Granted Jun 30, 2026

Inventors: Rong Zhao (Beijing, CN); Yun Zhang (Beijing, CN); Ya Wei Niu (Beijing, CN); Chen Li (Dezhou, CN); David Shao Chung Chen (Taipei City, TW)

Assignee: International Business Machines Corporation

G06F8/65

Abstract

Claims (36)

1 . A method comprising:

collecting data about a user's experience of a software application, wherein collecting data includes collecting input data including inputs submitted by the user from an input device on a device while the user is engaged with the software application and collecting visual data of the user while the user is engaged with the software application on the device;

processing the input data and the visual data, wherein eye-tracking data is extracted from the visual data and the eye-tracking data is processed using machine learning, wherein processing the input data and the visual data includes synchronizing the input data and the processed eye-tracking data based on time to obtain time-stamps of the input data and eye-tracking data;

analyzing, using machine learning, the processed input data and visual data to obtain usage patterns of the user by calculating a mean usage frequency over a first time period and by calculating a slope of usage frequency over the first time period;

analyzing, by a machine learning model trained on historical user experience data associated with the software application, the collected data to determine one or more suggested responses to one or more questions in a feedback form based on the usage patterns of the user;

determining, by the machine learning model trained on historical user experience data associated with the software application, a timing for presenting the feedback form to the user based on the slope of usage frequency as compared to the mean usage frequency over the first time period;

preparing the feedback form for the user based on the suggested responses;

presenting the prepared feedback form to the user at the determined timing; and

training the machine learning model based on an elapsed time the user spent filling out the feedback form and a number of options within the feedback form that the user modified.

2 . The method of claim 1 , wherein the preparing includes pre-populating the feedback form with the suggested responses.

3 . A computer system, the computer system comprising:

a processor set;

one or more computer readable storage media; and

program instructions stored on the one or more computer readable storage media to cause the processor set to perform operations comprising:

collecting data about a user's experience of a software application, wherein collecting data includes collecting input data including inputs submitted by the user from an input device on a device while the user is engaged with the software application and collecting visual data of the user while the user is engaged with the software application on the device;

processing the input data and the visual data, wherein eye-tracking data is extracted from the visual data and the eye-tracking data is processed using machine learning, wherein processing the input data and the visual data includes synchronizing the input data and the processed eye-tracking data based on time to obtain time-stamps of the input data and eye-tracking data;

analyzing, using machine learning, the processed input data and visual data to obtain usage patterns of the user by calculating a mean usage frequency over a first time period and by calculating a slope of usage frequency over the first time period;

analyzing, by a machine learning model trained on historical user experience data associated with the software application, the collected data to determine one or more suggested responses to one or more questions in a feedback form based on the usage patterns of the user;

determining, by the machine learning model trained on historical user experience data associated with the software application, a timing for presenting the feedback form to the user based on the slope of usage frequency as compared to the mean usage frequency over the first time period;

preparing the feedback form for the user based on the suggested responses;

presenting the prepared feedback form to the user at the determined timing; and

training the machine learning model based on an elapsed time the user spent filling out the feedback form and a number of options within the feedback form that the user modified.

4 . The computer system of claim 3 , wherein the preparing includes pre-populating the feedback form with the suggested responses.

5 . A device comprising:

a processor set;

one or more computer readable storage media; and

program instructions stored on the one or more computer readable storage media to cause the processor set to perform operations comprising:

collecting input data including inputs submitted by a user from a mouse and from a keyboard communicatively coupled to the device while the user is engaged with a software application;

collecting, from a webcam of the device, visual data of the user while the user is engaged with the software application on the device;

processing the input data and the visual data, wherein eye-tracking data is extracted from the visual data and the eye-tracking data is processed using machine learning, wherein processing the input data and the visual data includes synchronizing the input data and the processed eye-tracking data based on time to obtain time-stamps of the input data and eye-tracking data;

analyzing, using machine learning, the processed input data and visual data to obtain usage patterns of the user by calculating a mean usage frequency over a first time period and by calculating a slope of usage frequency over the first time period;

analyzing, by a machine learning model trained on historical user experience data associated with the software application, the collected data to determine one or more suggested responses to one or more questions in a feedback form based on the usage patterns of the user;

determining, by the machine learning model trained on historical user experience data associated with the software application, a timing for presenting the feedback form to the user based on the slope of usage frequency as compared to the mean usage frequency over the first time period;

preparing the feedback form for the user based on the suggested responses;

presenting the prepared feedback form to the user; and

training the machine learning model based on an elapsed time the user spent filling out the feedback form and a number of options within the feedback form that the user modified.

Assignments (1)

Continuity (1)

References Cited (25)