EP-4740161-A1 - METHOD AND SYSTEM FOR DETECTING FRAUDULENT TRANSACTIONS INVOLVING NON-FUNGIBLE TOKENS (NFT)

EP4740161A1EP 4740161 A1EP4740161 A1EP 4740161A1EP-4740161-A1

Abstract

A method for scoring authenticity of a non-fungible token (NFT) using multiple, disparate data sets includes: receiving a scoring request for the NFT; determining a marketplace authenticity score for a marketplace where the NFT is available for purchase based on marketplace metrics; determining a visual authenticity score based on a comparison of visual features of the NFT to visual features of trusted NFTs; determining a wallet authenticity score for a blockchain wallet associated with ownership of the NFT based on a transaction history for the blockchain wallet; calculating a confidence score for the NFT based on a combination of the marketplace, visual, and wallet authenticity scores, the confidence score representing a likelihood that the NFT is authentic; and transmitting the calculated confidence score in response to the received scoring request.

Inventors

ARORA, GARIMA
Patankar, Adarsh

Assignees

Mastercard International Incorporated

Dates

Publication Date: 20260513
Application Date: 20240617

Claims (18)

1. A method for scoring authenticity of a non-fungible token (NFT) using multiple, disparate data sets, comprising: receiving, by a receiver of a processing server, a scoring request for the NFT including at least one of the NFT or an identification value associated with the NFT; determining, by a processor of the processing server, a marketplace authenticity score for a marketplace where the NFT is available for purchase based on at least one or more marketplace metrics; determining, by the processor of the processing server, a visual authenticity score based on at least a comparison of one or more visual features of the NFT to visual features of a plurality of trusted NFTs; determining, by the processor of the processing server, a wallet authenticity score for a blockchain wallet associated with ownership of the NFT based on at least a transaction history for the blockchain wallet; calculating, by the processor of the processing server, a confidence score for the NFT based on a combination of at least the determined marketplace authenticity score, the visual authenticity score, and the wallet authenticity score, the confidence score representing a likelihood that the NFT is authentic; and transmitting, by a transmitter of the processing server, the calculated confidence score in response to the received scoring request.
2. The method of claim 1, wherein the one or more marketplace metrics includes at least one of: a popularity rank, domain registration data, network activity, and social media traffic for a webpage or application program associated with the marketplace.
3. The method of claim 1, wherein the marketplace authenticity score is determined using an Extreme Gradient Boosting model.
4. The method of claim 1, wherein the marketplace authenticity score is further based on a comparison of one or more visual features of the marketplace to visual features of a plurality of trusted marketplaces.
5. The method of claim 4, wherein the comparison of the one or more visual features of the marketplace to visual features of a plurality of trusted marketplaces uses one of: a Siamese model, a Deep Learning model, and Feature Maps.
6. The method of claim 1, wherein the marketplace authenticity score is further based on whether or not the marketplace is identified in a database of trusted marketplaces.
7. The method of claim 1, wherein the comparison of the one or more visual features of the NFT to visual features of a plurality of trusted NFTs uses one of: a Siamese model, a Deep Learning model, and Feature Maps.
8. The method of claim 1, further comprising: determining, by the processor of the processing server, a transaction authenticity score based on at least a transaction history for a transaction account associated with the NFT, wherein the confidence score is further based on the transaction authenticity score.
9. The method of claim 1, wherein the wallet authenticity score is further based on whether or not the blockchain wallet is identified in a database of suspicious blockchain wallets.
10. A system for scoring authenticity of a non-fungible token (NFT) using multiple, disparate data sets, comprising: a blockchain wallet associated with ownership of the NFT; a marketplace; and a processing server, the processing server including a receiver receiving a scoring request for the NFT including at least one of the NFT or an identification value associated with the NFT; a processor determining (i) a marketplace authenticity score for the marketplace where the NFT is available for purchase based on at least one or more marketplace metrics, (ii) a visual authenticity score based on at least a comparison of one or more visual features of the NFT to visual features of a plurality of trusted NFTs, and (iii) a wallet authenticity score for the blockchain wallet based on at least a transaction history for the blockchain wallet, and calculating a confidence score for the NFT based on a combination of at least the determined marketplace authenticity score, the visual authenticity score, and the wallet authenticity score, the confidence score representing a likelihood that the NFT is authentic; and a transmitter transmitting the calculated confidence score in response to the received scoring request.
11. The system of claim 10, wherein the one or more marketplace metrics includes at least one of: a popularity rank, domain registration data, network activity, and social media traffic for a webpage or application program associated with the marketplace.
12. The system of claim 10, wherein the marketplace authenticity score is determined using an Extreme Gradient Boosting model.
13. The system of claim 10, wherein the marketplace authenticity score is further based on a comparison of one or more visual features of the marketplace to visual features of a plurality of trusted marketplaces.
14. The system of claim 13, wherein the comparison of the one or more visual features of the marketplace to visual features of a plurality of trusted marketplaces uses one of: a Siamese model, a Deep Learning model, and Feature Maps.
15. The system of claim 10, wherein the marketplace authenticity score is further based on whether or not the marketplace is identified in a database of trusted marketplaces.
16. The system of claim 10, wherein the comparison of the one or more visual features of the NFT to visual features of a plurality of trusted NFTs uses one of: a Siamese model, a Deep Learning model, and Feature Maps.
17. The system of claim 10, wherein the processor of the processing server determines a transaction authenticity score based on at least a transaction history for a transaction account associated with the NFT, and the confidence score is further based on the transaction authenticity score.
18. The system of claim 10, wherein the wallet authenticity score is further based on whether or not the blockchain wallet is identified in a database of suspicious blockchain wallets.

Description

METHOD AND SYSTEM FOR DETECTING FRAUDULENT TRANSACTIONS INVOLVING NON-FUNGIB LE TOKENS (NFT) CROSS-REFERENCE TO RELATED APPLICATION This application claims the benefit of, and priority to, U.S. Patent Application No. 18/218,254, filed July 5, 2023. The entire disclosure of the above application is incorporated herein by reference. FIELD The present disclosure relates to the detection of potential fraud in cryptocurrency transactions, specifically the scoring of authenticity for non-fungible tokens (NFTs) using marketplace, visual, and transactional data. BACKGROUND Blockchains were first created as a way of providing for a cryptographic currency that could be transferred among participants in a decentralized manner that provided the participants with anonymity. Over time, participants discovered new uses for blockchains in a variety of different industries and applications. A recent new application for blockchains is in conjunction with non- fungible tokens, most commonly referred to as “NFTs.” An NFT is a unique digital object that can be bought and sold, whose provenance is tracked on a blockchain. At its inception, NFTs were most often digital artwork, but they have since expanded to also include other digital objects, such as representing items in online video games, songs, sports video clips, etc. An NFT, once created, is stored in a blockchain with transfers of ownership recorded therein. Much of the value of an NFT comes from its uniqueness; the purchaser can claim ownership of the NFT and show it off to others the same way an art collector can. Still, like with traditional, physical paintings, there is little to stop someone from creating a copy of the digital object. In most cases, the copy is marketed as such and has a significantly lower value than the original. However, nefarious actors can make a copy of an existing NFT, and present is as the original (e.g., representing that they have the real Mona Lisa), can make an NFT and claim to be a famous artist (e.g., Banksy), or can make an NFT that they claim comes from a famous artist (e.g., representing a painting having been done by Picasso). Currently, there are no systems designed for authenticating an NFT or the seller thereof. While a more tech savvy purchaser can identify the history of an NFT on a blockchain, they can still be unable to determine if the NFT itself or the seller are genuine. Thus, there is a need for a technical system that can provide a measure of authenticity for an NFT, regarding authenticity of the NFT itself as well as the marketplace on which the NFT is sold and the seller of the NFT, for users prior to purchase. SUMMARY The present disclosure provides a description of systems and methods for scoring the authenticity of a non-fungible token (NFT) using multiple, disparate data sets. A requestor can request the scoring of an NFT by a processing server. The processing server can gather data from multiple sources to determine authenticity scores for various aspects of the NFT and its sale. The processing server can determine a first score (herein referred to as a marketplace authenticity score) for the marketplace, such as a webpage or application program, where the NFT is sold, which can be based on marketplace metrics, such as network traffic, social media traffic, domain registration data, and popularity, and also on a comparison of the visuals of the marketplace, to represent how likely that the marketplace is genuine and not, for example, a phishing website. The processing server can also determine a second score (herein referred to as a visual authenticity score) for the NFT itself, by comparing one or more visual features of the NFT to existing NFTs, and in particular existing trusted NFTs, to determine if the NFT being sold is a copy of an already existing NFT. The processing server can also determine a third score (herein referred to as a wallet authenticity score) for the blockchain wallet that is selling the NFT based on at least its transaction history, which can represent a likelihood that the blockchain wallet is genuine and was the actual creator of the NFT or an authorized owner, and a likelihood that the blockchain wallet has not participated in past fraudulent transfers. The processing server can also determine a fourth score, which can be an additional risk score based on gathered marketplace and/or transactional data, such as if the marketplace supports buying and/or selling NFTs via the use of payment cards. The processing server can calculate a confidence score for the sale of the NFT using all of the authenticity scores, and provide the confidence score back to the requestor, who can then decide whether or not to purchase the NFT using the confidence score for guidance. The result makes for a significantly more well- informed purchaser regardless of technological savvy based on a variety of different data that can be prohibitively difficult for a nefarious actor to deceive. A method for scoring a