Using a machine learning model to optimize groupings in a breakout session in a virtual classroom
A computer-implemented method and a computer program product for optimizing groupings in a breakout session in a virtual classroom. A computer retrieves profiles of students in the virtual classroom, where the profiles of the students include data of academic performances of the students. The computer retrieves a known correspondents archive which includes historic data of productivities correlated to interactions among the students. In response to initialization of the breakout session in the virtual classroom, the computer determines optimal groups that yield most productive results, based on the profiles of the students, the known correspondents archive, and requirements of group settings given by an instructor, using a machine learning model. The computer provides the instructor with the optimal groups. In another embodiment, the computer analyzes context of customized assignments for the breakout session and determines the optimal groups further based on the context to the customized assignments.
1 . A computer-implemented method for optimizing groupings in a breakout session in a virtual classroom, the method comprising:
retrieving profiles of students in the virtual classroom, the profiles of the students including data of academic performances of the students;
retrieving a known correspondents archive which includes historic data of productivities correlated to interactions among the students, the known correspondents archive including productivity scores ranking pairs of the students;
detecting an initialization of the breakout session in the virtual classroom based upon an utterance of a specific phrase;
in response to detecting the initialization of the breakout session in the virtual classroom, determining optimal groups that yield most productive results, based on the profiles of the students, the known correspondents archive, and requirements of group settings given by an instructor, using a machine learning model;
providing the instructor with the optimal groups; and
training the machine learning model to select future groupings using at least one successful grouping as positive reinforcement input data and at least one unsuccessful grouping as negative reinforcement input data, wherein the at least one successful grouping and the at least one unsuccessful grouping is determined in part using a speech tone analyzer.
2 . The computer-implemented method of claim 1 , further comprising:
monitoring activities of the students in the breakout session; and
updating the profiles of the students and the known correspondents archive,
wherein the virtual classroom comprises a virtual conferencing platform.
3 . The computer-implemented method of claim 2 , further comprising:
receiving, from the instructor and the students, feedback on evaluations of the groupings in the breakout session; and
determining whether the groupings are successful by analyzing the results of the breakout session;
in response to determining that the groupings are successful, inputting the at least one successful groupings as the positive reinforcement input data into the machine learning model; and
in response to determining that the groupings are unsuccessful, inputting the at least one unsuccessful groupings as the negative reinforcement input data into the machine learning model and prompting the machine learning model to modify data and future groupings.
4 . The computer-implemented method of claim 1 , further comprising:
prompting the instructor and requiring the instructor to input the requirements of the group settings.
5 . The computer-implemented method of claim 4 , further comprising: receiving, from the instructor, a requirement of a size of each group.
6 . The computer-implemented method of claim 4 , further comprising:
receiving, from the instructor, a requirement about how the students are grouped; and
determining whether groups each including students with similar academic profiles or groups each including students with diverse academic profiles are requested.
7 . The computer-implemented method of claim 4 , further comprising:
receiving, from the instructor, a requirement about how the students are grouped; and
determining whether groups each including students with well-established relationships or groups each including students with no well-established relationships are requested.
8 . A computer program product for optimizing groupings in a breakout session in a virtual classroom, the computer program product comprising a computer readable storage medium having program instructions embodied therewith, the program instructions executable by one or more processors, the program instructions executable to:
retrieve profiles of students in the virtual classroom, the profiles of the students including data of academic performances of the students;
retrieve a known correspondents archive which includes historic data of productivities correlated to interactions among the students, the known correspondents archive including productivity scores ranking pairs of the students;
in response to initialization of the breakout session in the virtual classroom, determine optimal groups that yield most productive results, based on the profiles of the students, the known correspondents archive, and requirements of group settings given by an instructor, using a machine learning model;
provide the instructor with the optimal groups; and
train the machine learning model to select future groupings using at least one successful grouping as positive reinforcement input data and at least one unsuccessful grouping as negative reinforcement input data, wherein the at least one successful grouping and the at least one unsuccessful grouping is determined in part using a speech tone analyzer.
9 . The computer program product of claim 8 , further comprising the program instructions executable to:
monitor activities of the students in the breakout session; and
update the profiles of the students and the known correspondents archive.
10 . The computer program product of claim 9 , further comprising the program instructions executable to:
receive, from the instructor and the students, feedback on evaluations of the groupings in the breakout session; and
determine whether the groupings are successful by analyzing the results of the breakout session;
in response to determining that the groupings are successful, input the at least one successful groupings as the positive reinforcement input data into the machine learning model; and
in response to determining that the groupings are unsuccessful, input the at least one unsuccessful groupings as the negative reinforcement input data into the machine learning model and prompt the machine learning model to modify data and future groupings.
11 . The computer program product of claim 8 , further comprising the program instructions executable to: prompt the instructor and require the instructor to input the requirements of the group settings.
12 . The computer program product of claim 11 , further comprising the program instructions executable to: receive, from the instructor, a requirement of a size of each group.
13 . The computer program product of claim 11 , further comprising the program instructions executable to:
receive, from the instructor, a requirement about how the students are grouped; and
determine whether groups each including students with similar academic profiles or groups each including students with diverse academic profiles are requested.
14 . The computer program product of claim 11 , further comprising program instructions executable to:
receive, from the instructor, requirement about how the students are grouped; and
determine whether groups each including students with well-established relationships or groups each including students with no well-established relationships are requested.
15 . A computer-implemented method for optimizing groupings in a breakout session in a virtual classroom, the method comprising:
retrieving profiles of students in the virtual classroom, the profiles of the students including data of academic performances of the students;
retrieving a known correspondents archive which includes historic data of productivities correlated to interactions among the students, the known correspondents archive including productivity scores ranking pairs of the students;
receiving, from an instructor, predefined group settings, wherein one of the predefined group settings requires that each group includes students with a similar academic level;
analyzing context of customized assignments for the breakout session, wherein the customized assignments have different levels of academic challenges for different groups;
determining optimal groups that yield most productive results, based on academic levels included in the profiles of the students, the known correspondents archive, the context of the customized assignments, and the predefined group settings, using a machine learning model;
providing the instructor with the optimal groups for the breakout session in the virtual classroom;
assigning the customized assignments to respective ones of the optimal groups; and
training the machine learning model to select future groupings using at least one successful grouping as positive reinforcement input data and at least one unsuccessful grouping as negative reinforcement input data, wherein the at least one successful grouping and the at least one unsuccessful grouping is determined in part using a speech tone analyzer.
16 . The computer-implemented method of claim 15 , further comprising:
monitoring activities of the students in the breakout session; and
updating the profiles of the students and the known correspondents archive.
17 . The computer-implemented method of claim 16 , further comprising:
receiving user login credentials for the instructor and the students, wherein the login credentials comprise audio voice recognition;
receiving, from the instructor and the students, feedback on evaluations of the groupings in the breakout session; and
determining whether the groupings are successful by analyzing the results of the breakout session;
in response to determining that the groupings are successful, inputting the at least one successful groupings as the positive reinforcement input data into the machine learning model; and
in response to determining that the groupings are unsuccessful, inputting the at least one unsuccessful groupings as the negative reinforcement input data into the machine learning model and prompting the machine learning model to modify data and future groupings.
18 . The computer-implemented method of claim 15 , further comprising:
in receiving the predefined group settings, receiving, from the instructor, a predefined size of each group.
19 . The computer-implemented method of claim 15 , further comprising:
in receiving the predefined group settings, receiving, from the instructor, a requirement about how the students are grouped; and
determining whether groups each including students with well-established relationships or groups each including students with no well-established relationships are requested.
20 . The computer-implemented method of claim 15 , wherein the virtual classroom comprises a virtual conferencing platform.