Sharing and executing custom machine learning algorithms
An embodiment of the invention may include a method, computer program product, and computer system for managing a machine learning algorithm. The embodiment may include a computing device that distributes a first algorithm to a plurality of computing devices. The embodiment may include updates to the first algorithm, to create a second algorithm, by a first device of the plurality of computing devices. The first device is grouped with other devices in a first cluster of devices of the plurality of computing devices. The first cluster of computing devices comprises more than one computing device. Updating the first algorithm is performed based on information shared amongst the first cluster of devices.
1. A method for managing a machine learning algorithm without directly sharing Personally Identifiable Information (PII) data between devices including a plurality of computing devices, the method comprising:
receiving, by a first edge device of the plurality of computing devices, a first Artificial Intelligence (AI) model, wherein
the first AI model is received from a cloud device of the devices, and
the plurality of computing devices is grouped in a cluster based on one of: a geographical location of each computing device of the plurality of computing devices, or a user group in a game associated with each computing device of the plurality of computing devices;
generating, by the first edge device, one or more prediction results, using the first AI model;
receiving, by the first edge device, at least one user feedback for the one or more prediction results, wherein the at least one user feedback corresponds to modifications of the one or more prediction results based on preferences of a user of the first edge device;
updating, by the first edge device, the first AI model based on the PII data associated with the first edge device, non-PII data associated with the first edge device, and the received at least one user feedback, wherein the non-PII data indicates choices or decisions made by the user;
generating, by the first edge device, a second AI model based on the updating of the first AI model one or more hyperparameters associated with the second AI model are calculated, using a lazy learning technique, based upon the PII data, the non-PII data, and the received at least one user feedback;
sending, by the first edge device to a second edge device of the plurality of computing devices grouped in the cluster, the one or more hyperparameters, and the non-PII data, wherein a third AI model stored on the second edge device is updated using the one or more hyperparameters and the non-PII data; and
sending, by the first edge device to the cloud device, data including the one or more hyperparameters, and the non-PII data, wherein
the data is sent by masking of the PII data,
the masking of the PII data corresponds to exclusion of the PII data from the data sent by the first edge device to the cloud device, and
a fourth AI model stored on the cloud device is updated using the one or more hyperparameters and the non-PII data.
2. The method of claim 1 , wherein the first AI model is based on a K-nearest neighbor algorithm.
3. The method of claim 1 , wherein a timing of the sending of the one or more hyperparameters from the first edge device to the second edge device is based on an availability of the first edge device.
4. A computer program product for managing a machine learning algorithm without directly sharing Personally Identifiable Information (PII) data between devices including a plurality of computing devices, the computer program product comprising:
one or more computer readable storage media; and
program instructions collectively stored on the one or more computer readable storage media to perform operations comprising:
receiving, by a first edge device of the plurality of computing devices, a first Artificial Intelligence (AI) model, wherein
the first AI model is received from a cloud device of the devices, and
the plurality of computing devices is grouped in a cluster based on one of: a geographical location of each computing device of the plurality of computing devices, or a user group in a game associated with each computing device of the plurality of computing devices;
generating, by the first edge device, one or more prediction results using the first AI model;
receiving, by the first edge device, at least one user feedback for the one or more prediction results, wherein the at least one user feedback corresponds to modifications of the one or more prediction results based on preferences of a user of the first edge device;
updating, by the first edge device, the first AI model based on the PII data associated with the first edge device, non-PII data associated with the first edge device, and the received at least one user feedback, wherein the non-PII data indicates choices or decisions made by the user;
generating, by the first edge device, a second Al model based on the updating of the first AI model one or more hyperparameters associated with the second AI model are calculated, using a lazy learning technique, based upon the PII data, the non-PII data, and the received at least one user feedback;
sending, by the first edge device to a second edge device of the plurality of computing devices grouped in the cluster, the one or more hyperparameters, and the non-PII data, wherein a third AI model stored on the second edge device is updated using the one or more hyperparameters and the non-PII data; and
sending, by the first edge device to the cloud device, data including the one or more hyperparameters, and the non-PII data, wherein
the data is sent by masking of the PII data,
the masking of the PII data corresponds to exclusion of the PII data from the data sent by the first edge device to the cloud device, and
a fourth AI model stored on the cloud device is updated using the one or more hyperparameters and the non-PII data.
5. The computer program product of claim 4 , wherein the first algorithm AI model is based on a K-nearest neighbor algorithm.
6. The computer program product of claim 4 , wherein a timing of the sending of the one or more hyperparameters from the first edge device to the second edge device is based on an availability of the first edge device.
7. A computer system for management of a machine learning algorithm without directly sharing Personally Identifiable Information (PII) data between devices including a plurality of computing devices, the computer system comprising:
one or more processors;
one or more computer-readable storage devices; and
program instructions stored on at least one of the one or more computer-readable storage devices, the program instructions executable by at least one processor of the one or more processors to cause the at least one processor to:
receive, by a first edge device of the plurality of computing devices, a first Artificial Intelligence (AI) model, wherein
the first AI model is received from a cloud device of the devices, and
the plurality of computing devices is grouped in a cluster based on one of: a geographical location of each computing device of the plurality of computing devices, or a user group in a game associated with each computing device of the plurality of computing devices;
generate, by the first edge device, one or more prediction results, using the first AI model;
receive, by the first edge device, at least one user feedback for the one or more prediction results, wherein the at least one user feedback corresponds to modifications of the one or more prediction results based on preferences of a user of the first edge device;
update, by the first edge device, the first AI model based on the PII data associated with the first edge device, non-PII data associated with the first edge device, and the received at least one user feedback, wherein the non-PII data indicates choices or decisions made by the user;
generate, by the first edge device, a second AI model based on the update of the first AI model, wherein one or more hyperparameters associated with the second AI model are calculated, using a lazy learning technique, based upon the PII data, the non-PII data, and the received at least one user feedback;
send, by the first edge device to a second edge device of the plurality of computing devices grouped in the cluster, the one or more hyperparameters, and the non-PII data, wherein a third AI model stored on the second edge device is updated using the one or more hyperparameters and the non-PII data; and
send, by the first edge device to the cloud device, data including the one or more hyperparameters, and the non-PII data, wherein
the data is sent based on the PII data being masked,
the PII data being masked corresponds to exclusion of the PII data from the data sent by the first edge device to the cloud device, and
a fourth AI model stored on the cloud device is updated using the one or more hyperparameters and the non-PII data.
8. The computer system of claim 7 , wherein the first AI model is based on a K-nearest neighbor algorithm.
9. The computer system of claim 7 , wherein a timing at which the hyperparameters are sent from the first edge device to the second edge device is based on an availability of the first edge device.