Detection of Deepfakes in Financial Transactions Using Algorand Blockchain Consensus Mechanism
摘要
Algorand is a new cryptocurrency confirming transactions in a few minutes and scales to a large number of users. Even if some users are bad and the network is momentarily partitioned, Algorand assures that users never have contradictory views of confirmed transactions. Contrarily, current cryptocurrencies support momentary forks and demand a long time to confirm transactions with great confidence. Algorand uses a unique Byzantine Agreement (BA) technique based on Verifiable Random Functions (VRFs) to scale consensus to many users in order to reach an agreement among users on the upcoming round of transactions. In order to agree on the following set of transactions, this technique enables users to secretly verify if they have been chosen to engage in the BA and to include evidence of their selection in the network messages. In BA used in Algorand, other than their private keys, no other private values are stored, so users can be instantaneously replaced after sending the message. This protects selected individuals from targeted attacks after their identities are known and also helps to detect deepfakes. Experimental results have proven that the confirmed transactions are achieved within a minute and detect deepfake users in minimum time.