US 11,886,959 B2
System and method for self-healing in decentralized model building for machine learning using blockchain
Sathyanarayanan Manamohan, Chennai (IN); Krishnaprasad Lingadahalli Shastry, Bangalore (IN); and Vishesh Garg, Bangalore (IN)
Assigned to Hewlett Packard Enterprise Development LP, Spring, TX (US)
Filed by Hewlett Packard Enterprise Development LP, Houston, TX (US)
Filed on Feb. 21, 2019, as Appl. No. 16/282,098.
Prior Publication US 2020/0272934 A1, Aug. 27, 2020
Int. Cl. G06N 20/00 (2019.01)
CPC G06N 20/00 (2019.01) 20 Claims
OG exemplary drawing
 
1. A system of decentralized machine learning (ML) comprising:
a self-healing computer node of a blockchain network comprising a plurality of computer nodes, the self-healing computer node when recovering from a fault condition within the blockchain network being programmed to:
determine, by the self-healing computer node, that the self-healing computer node is unable to share training parameters in a first iteration of training a machine-learning model;
based on the determination, generate a first blockchain transaction comprising an indication that the self-healing computer node is out-of-sync with the first iteration of training the machine-learned model, wherein the first blockchain transaction is to be added to a distributed ledger and informs the plurality of computer nodes that the self-healing computer node is not ready to participate in a second iteration of training the machine-learned model;
obtain a global ML state from the distributed ledger;
compare the obtained global ML state with a local ML state at the self-healing computer node to determine whether the local ML state is consistent with the global ML state;
based on the comparing, determine that the global ML state is not consistent with the local ML state;
in response to the determining recover the local ML state using a trigger of a corrective action;
determine, by the self-healing computer node, that the self-healing computer node is able to share the training parameters in the second iteration of training the machine-learning model;
based on the determination, generate a second blockchain transaction comprising an indication that the self-healing computer node is in-sync with the second iteration of training the machine-learned model, wherein the second blockchain transaction is to be added to the distributed ledger and informs the plurality of computer nodes that the self-healing computer node is ready to participate in the second iteration of training the machine-learned model; and
re-enroll the self-healing computer node with the blockchain network to participate in the second iteration of training the machine-learned model.