Systems and methods for monetizing data in decentralized model building for machine learning using a blockchain
AssigneeHewlett Packard Enterprise Development LP
InventorsSATHYANARAYANAN MANAMOHAN, Vishesh Garg, Krishnaprasad Lingadahalli Shastry
Life of this patent
Estimating the term…
Abstract
Systems and methods are provided for leveraging blockchain technology in a swarm learning context, where nodes of a blockchain network that contribute data to training a machine learning model using their own local data can be rewarded. In order to conduct such data monetization in a fair and accurate manner, the systems and methods rely on various phases in which Merkle trees are used and corresponding Merkle roots are registered in a blockchain ledger. Moreover, any claims for a reward are challenged by peer nodes before the reward is distributed.
Claims20
A data monetization network node operating in a blockchain network, comprising:
- a processor;
- and
- a memory unit operatively connected to the processor including computer code that when executed causes the processor to:
- encrypt a raw data record extracted from a machine learning model being trained;
- create a secure cryptographic hash of the encrypted raw data record;
- build a hash tree based at least on the secure cryptographic hash of the encrypted raw data record, and register a corresponding hash tree root in a distributed ledger of the blockchain network;
- submit a claim for a reward represented by a height of the hash tree;
- provide a hash tree proof to each of a plurality of other network nodes in the blockchain network, the hash tree proof verifying the height of the hash tree in response to verification challenges from each of the plurality of other network nodes;
- calculate an amount of individual data points contributed by the data monetization network node based upon the height of the hash tree, the calculated amount of individual data points comprising a share of a monetization reward for training of the machine learning model;
- and
- receive the share of the monetization reward.
A data monetization network node operating in a blockchain network, comprising:
- a processor;
- and
- a memory unit operatively connected to the processor including computer code that when executed causes the processor to:
- encrypt a raw data record extracted from a machine learning model being trained;
- create and store a non-duplicative secure cryptographic hash of the encrypted raw data record;
- build a hash tree based at least on the non-duplicative secure cryptographic hash of the encrypted raw data record, and register a corresponding hash tree root in a distributed ledger of the blockchain network;
- submit a claim for a reward represented by a height of the hash tree;
- prove the height of the hash tree to each of a plurality of other network nodes in the blockchain network;
- calculate a percentage share of a monetization reward for training of the machine learning model;
- register the calculated percentage share of the monetization reward to the distributed ledger;
- and
- upon verification of the calculated percentage share of the monetization reward, receive the calculated percentage share of the monetization reward.
Dependencies are inferred from the claim text.