US 12,131,256 B2
System and method for training non-parametric machine learning model instances in a collaborative manner
Sathyanarayanan Manamohan, Karnatka (IN); Patrick Leon Gartenbach, Baden-Wurttemberg (DE); Markus Philipp Wuest, Baden-Wurttemberg (DE); Krishnaprasad Lingadahalli Shastry, Karnataka (IN); and Suresh Soundararajan, Karnataka (IN)
Assigned to Hewlett Packard Enterprise Development LP, Spring, TX (US)
Filed by Hewlett Packard Enterprise Development LP, Houston, TX (US)
Filed on Apr. 22, 2021, as Appl. No. 17/237,574.
Prior Publication US 2022/0215245 A1, Jul. 7, 2022
Int. Cl. G06N 3/08 (2023.01); G06N 3/04 (2023.01)
CPC G06N 3/08 (2013.01) [G06N 3/04 (2013.01)] 20 Claims
OG exemplary drawing
 
11. A blockchain system for training non-parametric Machine Learning (ML) model instances comprising:
a plurality of data processing nodes comprising:
a first subset of a plurality of data processing nodes, and
a second subset of the plurality of data processing nodes,
wherein the second subset is distinct from the first subset, wherein the first subset comprises more than one of a first processing node, the second subset comprises more than one of a second processing node, and the second processing node is different from the first processing node, and
wherein the first subset and the second subset comprise less than all of the plurality of data processing nodes,
wherein the first subset storing programmed instructions to:
train a non-parametric Machine Learning (ML) model instance over a first respective local dataset at the first subset of the plurality of data processing nodes to obtain a plurality of non-parametric ML model instances;
share the plurality of non-parametric ML model instances that are developed at the first subset of the plurality of data processing nodes with the second subset of the plurality of data processing nodes;
wherein the second subset storing programmed instructions to:
generate, at the second subset of the plurality of data processing nodes, a set of composite models at the second subset of the plurality of data processing nodes by processing the plurality of non-parametric ML model instances through a trainable parametric combinator;
train the set of composite models at the second subset of the plurality of data processing nodes, the training utilizing a second respective local dataset at the second subset of the plurality of data processing nodes in a coordinated learning technique to obtain a set of trained composite models,
wherein each of the second subset of the plurality of data processing nodes independently utilizes swarm learning to train the set of composite models; and
initiate an inference process of the set of trained composite models.