Updating of statistical sets for decentralized distributed training of a machine learning model

A computer-implemented method, comprising:

• selecting, by a system operatively coupled to a processor, using a randomization pattern, an additional statistical set associated with an additional machine learning node of a group of machine learning nodes from a group of statistical sets associated with respective machine learning nodes of the group of machine learning nodes, wherein the respective machine learning nodes are located at respective systems of a group of distributed systems comprising the system;
• and
• averaging, by a system operatively coupled to a processor, a statistical set associated with a machine learning node of the system with the additional statistical set to thereby compute an averaged statistical set during a training of the machine learning node.

A computer program product facilitating a process to compute an averaged statistical set, the computer program product comprising a computer readable storage medium having program instructions embodied therewith, the program instructions executable by a processor of a system to cause the processor to:

• selecting, by the processor, using a randomization pattern, an additional statistical set associated with an additional machine learning node of a group of machine learning nodes from a group of statistical sets associated with respective machine learning nodes of the group of machine learning nodes, wherein the respective machine learning nodes are located at respective systems of a group of distributed systems comprising the system;
• and
• average, by the processor, a statistical set associated with a machine learning node of the system with the additional statistical set to thereby compute an averaged statistical set during a training of the machine learning node.