| CPC G06Q 30/0218 (2013.01) [G06N 5/022 (2013.01); G06Q 30/0282 (2013.01)] | 36 Claims |

|
1. A method of assuring the validity of online product reviews of a product for online purchase, comprising:
executing, by at least one smart contract processed by at least one processing device, transactions on a blockchain implemented on a node, the transactions including a transaction that stores data relating to changes to online product review data stored on the blockchain, a transaction that stores an address of a crypto-wallet or application of a product reviewer, and a transaction created to record a product review submitted by the product reviewer, the at least one smart contract further specifying a number of tokens to be rewarded to the product reviewer and human validators for providing a valid product review, the at least one smart contract further including a state machine smart contract that stores key-value pairs to the blockchain where the keys represent different states of a state machine and the values represent corresponding actions or conditions that need to be fulfilled in order to transition to a next state of the state machine whereby the state machine manages a state of a product review on the blockchain;
annotating, using a machine learning (ML) annotator implemented by the at least one processing device, the product review from the product reviewer to generate an ML annotated review and providing a first score representative of a validity of the product review;
providing, by a workflow application implemented by the at least one processing device, the ML annotated review to a plurality of human validators for validation, wherein the plurality of human validators have staked at least one token for the ML annotated review;
enabling the plurality of human validators to use the workflow application to access ML annotated reviews and data associated with the ML annotated reviews for which each human validator has staked at least one token for at least one of the ML annotated reviews, to submit product reviews for a product, to fetch data and execute transactions on the blockchain, and to provide a second score representative of the validity of the ML annotated product review;
once a predetermined number of second scores have been collected from the human validators, categorizing, by an assured scoring engine implemented by the at least one processing device, each product review as valid or invalid based on a final score that is a predetermined function of the first score and the second scores;
rewarding, by the at least one smart contract implemented by the at least one processing device, tokens to the crypto-wallet or application of the product reviewer and at least one of the human validators for providing the product review based on the number of tokens to be rewarded for the valid product review by the at least one smart contract, the number of tokens staked against the product review in a proof-of-stake consensus model by the at least one of the human validators, and an accuracy of the product review, wherein the at least one smart contract receives tokens from at least one of the product review or human validators to fund the online product reviews;
providing, by the workflow application implemented by the at least one processing device, to at least one of the product reviewer or an owner of a product to which the product review is directed an indication of whether the product review is valid or invalid; and
storing, by the at least one smart contract, data relating to a status of the product review to the blockchain, wherein the status of the product review includes at least one of a current state of an online product review, validators assigned to the online product review, a current score of the online product review, annotations on the online product review, or a score produced by the ML annotator.
|