KR102096654B1 - Apparatus and method for approving instant deposit of cryptocurrency - Google Patents

Abstract

The present invention relates to a device for approving an immediate deposit of a cryptocurrency exchange using a dynamic fee, and a method thereof. The device for approving an immediate deposit of a cryptocurrency exchange comprises: a rejection probability prediction module including an artificial intelligence program code which predicts a rejection probability of an exchange deposit transaction based on transaction information of the exchange deposit transaction to generate rejection probability information; a dynamic fee calculation module generating an immediate deposit fee which is an expected loss which is dynamically calculated based on the rejection probability information generated by the rejection probability prediction module; and a deposit approval module determining whether a deposit for the exchange deposit transaction is approved based on approval of a user account for the immediate deposit fee.

Description

Apparatus and method for approving instant deposit of cryptocurrency for cryptocurrency exchanges using reinforced learning

The present invention relates to an apparatus and method for instantly depositing and accepting a cryptocurrency exchange using reinforcement learning.

Bitcoin, a cryptocurrency and cryptocurrency that appeared in the world in 2009, not only showed high demand in the market by shaking the definition of currency, but also had a ripple effect of applying blockchain to existing services. Satoshi Nakamoto, the founder of Bitcoin, proposed a method of combining cryptography and network technology to record and verify currency issuance and transaction history on a peer-to-peer distributed network at the same time, which is a distributed ledger called Blockchain ( distributed ledger) technology. Bitcoin is the first use case to demonstrate that blockchain technology actually works. The market cap of global cryptocurrency (virtual currency) reached $ 150B in August 2017. In addition, in August 2017, Bitcoin achieved a global daily trading volume of KRW 2,600 billion, Ethereum of KRW 900 billion, and Ripple of KRW 1.8 trillion. As of November 2018, there are more than 1,000 cryptocurrencies being traded.

Cryptocurrency exchange refers to a market place that trades cryptocurrency held by a specific user with fiat currency or other cryptocurrency. Currently, representative cryptocurrency exchanges include Bitumb, Upbit, Kobit, Coinone, and Upxide in Korea, and bitFlyer, Quoine, Coincheck in Japan, etc. There are BTCChina, OKCoin, gate.io, KuCoin of China and Hong Kong, bitMEX, Bittrex, Coinbase, Kraken, deyMEX of USA, Binance, Bitfinex, Bitstamp, HitBTC of Europe. Cryptocurrency exchanges are on the rise worldwide, and as of March 2017, you can trade cryptocurrencies in 152 exchanges in 76 currencies.

In these cryptocurrency exchanges, when performing cryptocurrency deposits and withdrawals such as coins and tokens, cryptocurrency deposits and withdrawals are performed based on the confirmation of transactions that include information about the state change between the user account address of the public blockchain and the account address of the exchange. . At this time, according to the creation or distribution of the transaction, the user does not immediately perform the deposit and withdrawal of cryptocurrency, but since the withdrawal of the cryptocurrency is performed based on the confirmation of the transaction, there is a problem that the user's cryptocurrency deposit and withdrawal must be delayed considerably. . This is because the stability of a transaction improves as blocks accumulate after the creation and distribution of a specific transaction.

The transaction confirmation means, for example, in order to transfer the bitcoin of the A address to the B address, it creates a transfer transaction (Transaction) and signs it with its own 'private key' to propagate it to the network. When the transaction is stored in the queue of each node's memory, the stored transaction is propagated after the miner receives it, inserts it into the block it mines and issues, and then propagates it. As such, if a new block is additionally formed after the transaction is confirmed and connected to the previous blockchain with a blockhead hash value, the number of confirmations becomes 2 or more. Each cryptocurrency exchange pre-sets the number of such confirmations and approves or confirms the deposit.

Accordingly, for example, the coin deposit and withdrawal from the A exchange to the B exchange proceeds to the next step. The first stage is the withdrawal approval step of the A exchange, and in order to transfer the coin of the A exchange to the B exchange, the withdrawal approval of the A exchange must first be performed. In exchange A, if financial fraud such as voice phishing is suspected, exchange A will refuse to approve the withdrawal. The second stage is the blockchain network approval (confirmation) stage, and when the withdrawal approval is completed at the A exchange, the coin is sent to the blockchain network to receive confirmation. At this time, if the fee to be paid (GAS in the case of Ethereum) is low, the approval may be delayed or canceled if the miner evades approval. The third step is an approval step for depositing B exchanges, and approval is automatically performed when the number of confirmations for deposit approval is satisfied at the exchange to be deposited. The number of confirmations required for approval is different for each coin and exchange.

Korea registered patent 10-1862000, Risk management system of cryptography money using multi-customer, Pop Funding Co., Ltd. Republic of Korea Patent Publication 10-2018-0101355, System and method for improving security in the exchange of blockchain assets (SYSTEMS AND METHODS FOR IMPROVING SECURITY IN BLOCKCHAIN-ASSET EXCHANGE), Shapeshift Age

However, according to the existing payment method of the cryptocurrency exchange using the confirmation of the transaction, there arises a problem that it takes too much time to deposit the cryptocurrency. This problem causes a huge loss, especially for investors who perform cryptocurrency day trading, financial transactions, scalping, and swing trading.

Accordingly, an object of the present invention is to provide an apparatus and method for approving deposits of cryptocurrency exchanges using dynamic fees to improve the speed of deposit approval.

Hereinafter, specific means for achieving the object of the present invention will be described.

An object of the present invention is to update the priority of a plurality of nodes constituting a blockchain distributed network and to store a program code configured to approve a deposit request of a specific user; And a processing module that processes the program code to determine whether to approve the deposit request, and the program code collects node information that is information of the plurality of nodes in the blockchain distributed network, and collects the node information. A node priority update step of generating node evaluation information for each node on the basis and updating the node priorities for the plurality of nodes using the node evaluation information; A transaction information request step of requesting transaction information, which is information on the transaction, including the number of confirmations of a transaction corresponding to the deposit request at an upper node that is a node having a high priority according to the updated node priority; A transaction information receiving step of receiving the transaction information from the upper node; And a transaction confirmation and deposit approval step of performing a deposit approval for the deposit request when the number of confirmations of the transaction in the transaction information reaches a predetermined number of confirmations as an approval condition. The node evaluation information includes update information, which is a rate of update response rate of each node with respect to an existing transaction or a transaction that satisfies the approval condition among transactions distributed by each node to the blockchain distributed network. Characterized in that it can be achieved, it can be achieved by providing a cryptocurrency exchange deposit approval device using artificial intelligence.

Another object of the present invention, after the cryptocurrency exchange deposit approval device collects node information, which is information of a plurality of nodes in the blockchain distributed network, and generates node evaluation information for each node based on the node information, A node priority update step of updating node priorities for the plurality of nodes using the node evaluation information; Transaction information, which is information about the transaction, including the number of times the transaction corresponding to the user's deposit request at a higher node, which is a node having a higher priority, according to the updated node priority, by the cryptocurrency exchange deposit approval device Requesting transaction information to request; Cryptocurrency exchange deposit approval device, the transaction information receiving step of receiving the transaction information from the upper node; And a transaction confirmation and deposit approval step of the cryptocurrency exchange deposit approval device performing deposit approval for the deposit request when the number of confirmations of the transaction in the transaction information reaches a preset number of confirmations as an approval condition. It is configured to be performed on a computer, and the node evaluation information includes an update latency of each node with respect to an existing transaction or a ratio of transactions satisfying the approval condition among transactions distributed by the nodes to the blockchain distributed network. It can be achieved by providing a method for approving deposits of cryptocurrency exchanges using artificial intelligence, characterized in that in-reputation information is included.

Another object of the present invention is to specify the transaction information of an exchange deposit transaction, which is a transaction that transfers a specific cryptocurrency from a user account that is a user account in a specific blockchain to an exchange account that is an account of an exchange in the specific blockchain. A transaction information collection module received from a connection node that is at least one node among a plurality of nodes constituting the distributed network of the blockchain; A rejection probability prediction module including an artificial intelligence program code for predicting rejectability of the exchange deposit transaction based on the transaction information of the exchange deposit transaction and generating rejection probability information; And a deposit approval module for determining whether to approve the deposit for the exchange deposit transaction based on the rejectability information; and the transaction information includes fee information set by the user account in the exchange deposit transaction, When the deposit for the exchange deposit transaction is approved by the deposit approval module, before the proof by the mining node of the block in which the exchange deposit transaction is recorded is performed, the user is within the exchange of the specific cryptocurrency approved. It can be achieved by providing an apparatus for accepting deposits of cryptocurrency exchanges using artificial intelligence, which is characterized in that the sale in is possible.

In addition, the AI program code of the rejection probability prediction module may reject the exchange deposit transaction based on the transaction information of the exchange deposit transaction and the transaction pool information of the connection node. To generate the rejection possibility information by predicting, the transaction pool information may include at least one of the number of transactions included in the transaction pool, the commission rank of the exchange deposit transaction, and the total of fees of the transaction pool.

Another object of the present invention, the transaction information collection module, the exchange deposit transaction, which is a transaction that transfers a specific cryptocurrency from a user account that is a user's account in a specific blockchain to an exchange account that is an account in the specific blockchain of the exchange. A transaction information collection step of receiving transaction information of a connection node which is at least one node among a plurality of nodes constituting the distributed network of the specific blockchain; A rejection predictability module including an artificial intelligence program code for predicting rejectability of the exchange deposit transaction based on the transaction information of the exchange deposit transaction and generating rejection probability information, performs the artificial intelligence program code Predicting the possibility of rejection; And a deposit approval step, wherein the deposit approval module determines whether to approve the deposit for the exchange deposit transaction based on the rejectability information; and the transaction information includes a fee set by the user account in the exchange deposit transaction. Information, and if the deposit for the exchange deposit transaction is approved by the deposit approval module, before the proof by the mining node of the block in which the exchange deposit transaction is recorded, the user is granted the specific password It can be achieved by providing a method for approving deposit of a cryptocurrency exchange using artificial intelligence, characterized in that the sale of the currency is possible within the exchange.

In addition, the AI program code of the rejection probability prediction module may reject the exchange deposit transaction based on the transaction information of the exchange deposit transaction and the transaction pool information of the connection node. To generate the rejection possibility information by predicting, the transaction pool information may include at least one of the number of transactions included in the transaction pool, the commission rank of the exchange deposit transaction, and the total of fees of the transaction pool.

In addition, the artificial intelligence program code of the rejectability predictive module includes attributes configuring transaction information of a specific transaction (the number of transactions sent from the user account, the fee information, the amount of the specific cryptocurrency to be transferred, and the size of data. ), The attributes that make up the block information (number of blocks, number of block lists in uncle blocks, sum of fees, difficulty and proof constants of blocks), transaction pool, which is the information in transaction pool, which is a set of transactions stored in memory before being written to the block. Initial generation by selecting a plurality of attribute groups that are attribute combinations that include at least some of the attributes constituting the information (the number of transactions included in the transaction pool, the commission rank of the exchange deposit transaction, and the sum of the fees in the transaction pool) Generating attribute group generating step; Input layer of the artificial neural network which samples some of the transactions in the past into a training set and outputs the transaction information, the block information, and the transaction pool information based on the plurality of attribute groups, the rejectability information for the specific transaction A learning step of learning the artificial neural network based on a prediction degree that is a distance between the rejection possibility information output from the output layer and the actual rejection of the specific transaction, input to the input layer; The artificial neural network for sampling the remaining part of the past transactions into a test set and outputting the transaction information, the block information, and the transaction pool information based on the plurality of attribute groups, the rejectability information for the specific transaction. Performance evaluation for evaluating the performance of the artificial neural network based on a prediction degree, which is a distance between the rejectability information input from the output layer and output from the output layer and whether the specific transaction is actually rejected or not. step; And an evolution step of selectively replacing a plurality of attribute groups based on the predictive degree of the performance evaluation step to generate a replacement generation consisting of the plurality of attribute groups selectively evolved. The attribute group is updated by repeating the input step or the evolution step until a specific termination condition is satisfied for the replacement generation, and the rejection probability prediction module is configured to replace the substitution after the specific termination condition is satisfied. It may be characterized in that the rejectability information of the exchange deposit transaction is generated by using the attribute group in which the highest prediction degree among the plurality of attribute groups constituting a household is output.

In addition, the artificial intelligence program code of the rejectability predictive module includes attributes configuring transaction information of a specific transaction (the number of transactions sent from the user account, the fee information, the amount of the specific cryptocurrency to be transferred, and the size of data. ), The properties that make up the block information (number of blocks, number of block lists in uncle blocks, sum of fees, difficulty and proof constants of blocks) and transaction pool, which is information from transaction pool, which is a set of transactions stored in memory before being written to the block. The attributes constituting the information (the number of transactions included in the transaction pool, the commission rank of the exchange deposit transaction, and the sum of the fees in the transaction pool) are set as an environment, and the rejection probability prediction module is set as an agent. A reinforcement learning model for generating a reinforcement learning model in which the output of the rejectability information of the rejectability prediction module for the specific transaction is an action and the actual rejection of the specific transaction is rewarded. Creation step; An action step in which the rejection probability prediction module, which is the agent, predicts a rejection possibility for a transaction in the past and outputs the rejection probability information; A reward step of performing the reward on the rejection probability prediction module, which is the agent, based on whether the actual transaction was rejected in the past; And an update step of updating the rejection probability prediction module based on the reward, performed on the computer, and repeating the action step or the update step until a specific termination condition is satisfied to predict the rejection probability. The module is updated, and the rejection probability prediction module may generate the rejection probability information of the exchange deposit transaction after the specific closing condition is satisfied.

Another object of the present invention is to specify the transaction information of an exchange deposit transaction, which is a transaction that transfers a specific cryptocurrency from a user account that is a user account in a specific blockchain to an exchange account that is an account of an exchange in the specific blockchain. A transaction information collection module received from a connection node that is at least one node among a plurality of nodes constituting the distributed network of the blockchain; A rejection probability prediction module including an artificial intelligence program code for predicting rejectability of the exchange deposit transaction based on the transaction information of the exchange deposit transaction and generating rejection probability information; A dynamic fee calculation module that generates an instant deposit fee, which is an expected loss dynamically calculated based on the rejection probability information generated by the rejection probability prediction module; And a deposit approval module for determining whether to approve the deposit for the exchange deposit transaction based on the approval of the user account for the immediate deposit fee, wherein the transaction information is set by the user account in the exchange deposit transaction If the deposit for the exchange deposit transaction is approved by the deposit approval module, including the fee information, before the proof by the mining node of the current block that is the block in which the exchange deposit transaction is recorded, the user It can be achieved by providing a cryptocurrency exchange deposit approval device using a dynamic fee, characterized in that it is possible to sell the approved cryptocurrency in the exchange.

In addition, the attributes constituting the transaction information of the exchange deposit transaction (the number of transactions sent from the user account, the fee information, the amount of the specific cryptocurrency to be remitted and the size of data), the attributes constituting the block information (of the block Number, number of block list of uncle block, sum of fees, difficulty and proof constant of block), and attributes to configure transaction pool information, which is information of transaction pool, which is a set of transactions stored in memory before being written to the block (included in the transaction pool) Confirm the time by which the exchange deposit transaction is recorded in a new block and confirmed by the mining node's block proof based on at least a part of the number of transactions made, the commission rank of the exchange deposit transaction, and the sum of the fees in the transaction pool) Further comprising a confirmation time prediction module for generating time prediction information, the dynamic fee calculation module, the rejectability information, the confirmation time prediction information predicted by the confirmation time prediction module and the transaction information of the specific cryptocurrency ( It may be characterized by generating an instant deposit fee, which is an expected loss that is dynamically calculated based on at least one of transaction volume, unit price per hour, and hourly price.

As described above, according to the present invention, there are the following effects.

First, according to an embodiment of the present invention, there is an effect of increasing the speed of depositing coins (cryptocurrency, virtual currency) of the cryptocurrency exchange.

The following drawings attached to the present specification illustrate preferred embodiments of the present invention, and serve to further understand the technical idea of the present invention together with the detailed description of the present invention, so the present invention is limited only to those described in those drawings. And should not be interpreted.
1 is a view showing the relationship of the cryptocurrency exchange deposit approval device according to an embodiment of the present invention,
Figure 2 is a flow chart showing a method of creating a deposit account of the cryptocurrency exchange deposit approval device 1 according to an embodiment of the present invention,
Figure 3 is a flow chart showing a method of approving the deposit of the cryptocurrency exchange deposit approval device 1 according to an embodiment of the present invention,
Figure 4 is a schematic diagram showing a specific configuration of the cryptocurrency exchange deposit approval device 1 according to an embodiment of the present invention,
5 is a schematic diagram showing a specific configuration of the cryptocurrency exchange deposit approval device 10 using artificial intelligence according to an embodiment of the present invention.

Hereinafter with reference to the accompanying drawings will be described in detail an embodiment that can be easily carried out by the person of ordinary skill in the art. However, in the detailed description of the operation principle of the preferred embodiment of the present invention, if it is determined that a detailed description of a related known function or configuration may unnecessarily obscure the subject matter of the present invention, the detailed description will be omitted.

In addition, the same reference numerals are used for parts having similar functions and functions throughout the drawings. Throughout the specification, when a specific part is said to be connected to another part, this includes not only the case of being directly connected, but also the case of being connected indirectly with another element in between. In addition, the inclusion of specific components does not exclude other components unless specifically stated otherwise, and means that other components may be further included.

Cryptocurrency exchange deposit approval device

1 is a view showing the relationship of the cryptocurrency exchange deposit approval device according to an embodiment of the present invention. As shown in Figure 1, the cryptocurrency exchange deposit approval device 1 according to an embodiment of the present invention is connected to the user client 2 and the blockchain distributed network, the block of the deposit request from the user client 2 It can be configured to confirm the on-chain confirmation and approve the deposit.

The user client 2 refers to a personal client such as a mobile device, desktop, or tablet PC where the user sends a request such as a deposit request or a deposit history request to the cryptocurrency exchange, and receives the deposit history.

The blockchain distributed network means a distributed network composed of a plurality of nodes 3 each including a blockchain 30 synchronized with each other. The node 3 communicates with the blockchain 30 stored in the memory and receives the state change of the blockchain 30 to update the blockchain 30 or other state change of the blockchain 30 in the blockchain distributed network. Refers to the configuration distributed to nodes.

The user client 2 sends a remittance request to a specific node, which is a request for remitting a specific number of a specific cryptocurrency from a user account to an exchange account, and the specific node that receives the remittance request creates a transaction to create another transaction in the blockchain distributed network. It is distributed to the node 3. The distributed transaction is stored in the memory of each node 3, and when a block is proved by a specific mining node, the transaction is included in the authenticated block and recorded in the blockchain 30. The cryptocurrency exchange deposit approval device 1 receives the information of the transaction through the node 3, checks the number of blocks after the block in which the transaction is verified, which is the number of confirmations for the transaction, and approves the deposit for the transaction Therefore, the user can process transactions using a specific number of cryptocurrencies as many as the number deposited in the cryptocurrency exchange.

With respect to the creation of the deposit account, FIG. 2 is a flow chart showing a method of creating a deposit account by the deposit acceptance device 1 for a cryptocurrency exchange according to an embodiment of the present invention. As shown in Figure 2, the cryptocurrency exchange deposit approval device 1 at the request of the user client 2 ① cryptocurrency deposit address request step, ② account creation request step, ③ account address and private key return step, ④Account address output step can be included.

① The request step of the cryptocurrency deposit address is a step in which the user client 2 requests the deposit address for a specific cryptocurrency from the cryptocurrency exchange deposit approval device 1.

② The account creation request step is a step in which the cryptocurrency exchange deposit approval device 1 requests the specific node 3 to create an account for the specific cryptocurrency.

③Account address and private key return step is to create an account for a specific cryptocurrency in node 3, and return the account address, which is the public key of the account, and the private key of the account to the cryptocurrency exchange deposit approval device 1 to be.

④The account address output step is a step of transmitting the generated account address, which is a deposit address for a specific cryptocurrency, to the user client 2 in the cryptocurrency exchange deposit approval device 1 and outputting it. The user may request a specific node 3 to use a cryptocurrency exchange by remitting a specific number of cryptocurrencies to an account address output to the user client 2.

In connection with the deposit approval, FIG. 3 is a flowchart illustrating a method for approving the deposit of the apparatus 1 for approving deposit of a cryptocurrency exchange according to an embodiment of the present invention. As illustrated in FIG. 3, the method of depositing the payment of the cryptocurrency exchange deposit approval device 1 includes: ① cryptocurrency transfer request step, ② transaction creation and distribution step, ③ node priority update step, ④ transaction information request step, It may include ⑤transaction information transmission step, ⑥transaction confirmation and deposit approval step.

① In the step of requesting the remittance of cryptocurrency, the user client 2 sends a remittance from the user account to the exchange account address, which is the deposit address, for a specific number of specific cryptocurrencies to the first node, one of the nodes 3 of the blockchain distributed network. This is the request phase.

② The transaction creation and distribution phase is performed by creating and distributing a transaction corresponding to the cryptocurrency transfer request on the first node, and then creating a new block on the blockchain 30 and specifying it as one of the nodes 3 of the blockchain distributed network. When the mining node proves a new block of the blockchain 30, it is a step of distributing the transaction to the node 3 of the entire blockchain distributed network by including the transaction in the block. The distributed block is verified by each node 3, and when adequacy is verified, a new block containing the transaction is linked to the existing blockchain 30 as a linked list.

③ In the node priority update step, the cryptocurrency exchange deposit approval device 1 collects information of a plurality of nodes 3 of the blockchain distributed network, generates node evaluation information, and then uses the node evaluation information to evaluate the node priority. It is a step to update. In the node evaluation information according to an embodiment of the present invention, the update response rate of each node with respect to the existing transaction, the reputation information of each node (the ratio of transactions that satisfy the approval condition among the distributed transactions) And the like.

④The transaction information request step is transaction information, which is information about a transaction, including the number of confirmations of the transaction at the second node (higher node), which is the node 3 having the high priority by the cryptocurrency exchange deposit approval device 1. It is a step to request. According to an embodiment of the present invention, the cryptocurrency exchange deposit approval device 1 may request transaction information from the second node at the expected confirmation time by predicting the expected prediction time of the transaction. According to an embodiment of the present invention, the expected confirmation time of a transaction may be calculated using existing transaction information of the second node.

⑤The transaction information transmission step is a step of transmitting the transaction information from the second node to the cryptocurrency exchange deposit approval device 1.

⑥ The transaction confirmation and deposit approval step is a step in which the cryptocurrency exchange deposit approval device 1 checks the number of confirmations of the transaction and, when the predetermined number of confirmations has been reached, performs a deposit approval for the transaction. When the deposit approval is performed for the transaction, the user can trade the cryptocurrency deposited in the cryptocurrency exchange. According to another embodiment of the present invention, the cryptocurrency exchange deposit approval device 1 checks the number of confirmations of the transaction, predicts the likelihood information that is the probability of reaching the predetermined number of confirmations, and based on the approval possibility information It is possible to decide whether to approve the deposit for the transaction. Prediction of the applicability information may be calculated based on information on the second node, information on the first node that is the transaction distribution node, and transaction information on the current second node. For example, the cryptocurrency exchange deposit approval device 1 may calculate the probability of being approved within a specific time for a transaction distributed in the first node and confirmed in the second node as approval possibility information.

With respect to the specific configuration of the cryptocurrency exchange deposit approval device 1, FIG. 4 is a schematic diagram showing the specific configuration of the cryptocurrency exchange deposit approval device 1 according to an embodiment of the present invention. 4, the cryptocurrency exchange deposit approval device 1 according to an embodiment of the present invention includes a node information collection module 100, a node evaluation information generation module 101, and a node priority update module 102 ), Tx information collection module 103, Tx confirmation module 104, and deposit approval module 105.

The node information collection module 100 is a module that collects node information, which is information of at least one node 3 constituting a blockchain distributed network including the first node and the second node. Node information may include distribution time of an existing transaction, reception time, block information for each transaction, and the like.

The node evaluation information generation module 101 is a module that generates node evaluation information for each node 3 based on the collected node information. In the node evaluation information, the update response speed of each node for the existing transaction (latency, the time from the transaction distribution of the first node to the reception of the node or the number of blocks generated), the replication information of each node (distribution) The percentage of transactions that satisfy the approval condition among one transaction) may be included.

The node priority update module 102 is a module that updates node priority information that is a priority for each node 3 based on the node evaluation information.

The Tx information collection module 103 is a module that collects transaction information that is Tx (Transaction) information about a deposit account requested by the user client 2 from the second node, which is a node within a priority set in the node priority information.

The Tx confirmation module 104 is a module that checks whether the transaction is confirmed by generating confirmation information, which is information on how many blocks are connected after the block in which the corresponding transaction is recorded on the blockchain.

The deposit approval module 105 is a module that approves the deposit of the user client 2 based on the confirmation information for the transaction.

Cryptocurrency exchange deposit approval device using artificial intelligence

According to another embodiment of the present invention, artificial intelligence may be used in a cryptocurrency exchange deposit device. With respect to the specific configuration of the cryptocurrency exchange deposit approval device 10 using artificial intelligence, FIG. 5 shows the specific configuration of the cryptocurrency exchange deposit approval device 10 using artificial intelligence according to an embodiment of the present invention. It is a schematic diagram. As shown in Figure 5, the cryptocurrency exchange deposit approval device 10 using artificial intelligence according to an embodiment of the present invention is a transaction information collection module 110, block information collection module 111, rejection probability prediction module (112), a confirmation time module 113, a dynamic fee calculation module 114, and may include a deposit approval module 115. The cryptocurrency exchange deposit approval device 10 using artificial intelligence according to an embodiment of the present invention is connected to one or more connection nodes 4 which are part of a plurality of nodes 3 constituting the entire blockchain distributed network, and transaction information And block information can be received or collected. At this time, the connection node 4 may be composed of a full node capable of mining, not a light weight node.

The transaction information collection module 110 is a module that collects transaction information that is information of a new transaction received by the connection node 4. In particular, the transaction information collection module 110 searches for an exchange deposit transaction using the received (to) among the received new transactions as an exchange account connected to the cryptocurrency exchange deposit approval device 10 using artificial intelligence, and collects transaction information (110) The transaction information collected for the exchange deposit transaction is, for example, in the case of an Ethereum-based blockchain, nonce (number of transactions sent from the account), gas price (cost to pay per gas), gas limit (Maximum amount of gas to be paid to execute the transaction), value (amount of cryptocurrency to be transferred), v, r, s (signature identifier), data (input data of message call), etc. . At this time, gas price and gas limit may be defined as fee information.

In addition, the transaction information collection module 110 may be configured to collect transaction pool information, which is information about the transaction pool of the connection node 4. The transaction pool means a set of transactions collected to be stored in a specific block before a new block is determined by mining by the mining node. The transaction information collection module 110 may be configured to collect transaction information (nonce, to, gas price, gas limit, value, v, r, s, data) for at least one transaction included in the transaction pool .

The block information collection module 111 is a module that collects block information that is information of a block proved by a miner node. For block information, for example, in the case of an Ethereum-based blockchain, Block Num (number of blocks), Prev Hash (block header hash value of parent blocks), Timestamp (unix timestamp of blocks), Uncles Hash (of Uncle Blocks) Block list hash value), Beneficiary (Miner's account address that proved the block), Gas Limit (maximum number of gas a block can contain), Gas Used (gas used for performing tx included in the block) Total), Difficulty (difficulty of the block), Nonce (provided that the block performed sufficient computation in combination with Mix hash), Mix hash (proven that the block performed sufficient computation in combination with Nonce), etc. have.

The rejection probability prediction module 112 rejects an exchange deposit transaction based on transaction information, transaction pool information, and block information of the exchange deposit transaction collected by the transaction information collection module 110 and the block information collection module 111. ) It is a module that predicts the possibility and generates rejection possibility information. The rejection probability prediction of the exchange deposit transaction of the rejection prediction module 112 may be performed by an artificial intelligence-based prediction model such as basic machine learning, artificial neural network, genetic algorithm, and SVM using supervised learning or unsupervised learning.

Regarding the prediction model based on the supervised learning of the rejectability prediction module 112, the rejected transaction information, which is the information of the rejected transaction that is not recorded as a block in the transaction pool among the existing transactions, is supervised learning of the prediction model Can be used to Whether or not to reject the transaction may be configured and learned as a dummy value, and the rejection probability information output by the rejection probability prediction module 112 may be output as a probability value.

Regarding the prediction model based on unsupervised learning of the rejection probability prediction module 112, the rejectability information is predicted by predicting the rejectability of exchange deposit transaction information through clustering of transaction information including exchange deposit transaction information and reject transaction information Can output The rejection probability prediction module 112 may generate rejection probability information in the next step.

(1) Vectorize by embedding existing transaction information, transaction pool information before the transaction is written to the block, and block information before the transaction is written to the block

(2) Clustering the vectorized information to determine the weight for each attribute that best distinguishes the rejected transaction group from the rejected transaction group

(3) The determined weight is applied to each attribute of the exchange deposit transaction

(4) Calculating which group is closer to the rejected transaction group or the unrejected transaction group to generate exchange rejection information

According to this, it is possible to accurately predict the possibility of rejection based on the data collected by the transaction information collection module 110 and the block information collection module 111.

With respect to the prediction model based on the genetic algorithm and the artificial neural network of the rejection probability prediction module 112, the rejection prediction module 112 may be updated to the next step and generate rejection information with improved accuracy.

(1) Attributes that make up the transaction information of the transaction (number of transactions sent from the account, fee information, amount of cryptocurrency to be transferred, size of data), and Attributes that make up the block information (number of blocks, blocks of uncle blocks Among the number of lists, the sum of fees, the difficulty of blocks, proof constants such as nonce and mix hash), and the attributes that make up the transaction pool information (the number of transactions in the transaction pool, the commission rank of a specific transaction, and the sum of the fees in the transaction pool) Generation of initial generation by selecting a plurality of attribute groups that are attribute combinations that include at least some attributes

(2) The input layer of the artificial neural network that sets transaction information, block information, and transaction pool information related to a part of past transactions as a training set based on a plurality of attribute groups of the corresponding household, and outputs rejectability information for a specific transaction. Learning the artificial neural network by back propagation based on the predictability, which is the distance between the rejection information output by inputting the training set in (Layer) and the actual rejection of the transaction.

(3) Transaction information, block information, and transaction pool information related to the rest of the past transactions based on a plurality of attribute groups of the household are set as a test set, and predictability (accuracy or AUC, when a test set is entered) Area under the curve) to evaluate the performance of the artificial neural network

(4) Selective evolution (selection, crossover, mutation) of a plurality of attribute groups of the corresponding generation based on the predictive diagram to generate a selectively evolved replacement generation

(5) Proceeding from step 2 to step 3 for the generated replacement generation to generate predictions for the replacement generation

(6) When the predicted degree of the generated replacement generation satisfies a specific termination condition, the selective evolution is terminated, and when the generated prediction degree does not satisfy the specific termination condition, the next generation is generated by performing selective evolution in 4 steps (or In this case, the termination condition may be the case where the attribute groups of the generated replacement generations are almost identical to each other, or the change in the generation of dashes is below a certain level.)

(7) After the selection evolution is over, information on the likelihood of rejecting the newly entered exchange deposit transaction is output using the most predictable attribute group among the plurality of attribute groups of the last generation of substitution.

In the above, the genetic algorithm genotype is each attribute, chromosome is each attribute group, generation is a set of attribute groups of a specific stage, which is a plurality of chromosomes, parent generation is the start of selective evolution, child generation is newly created by selective evolution It can mean a set of attribute groups. The length of the chromosome may be composed of the same number as the node of the input layer of the artificial neural network, which is less than the total number of attributes constituting the transaction information of the exchange deposit transaction, the attributes constituting the block information, and the attributes constituting the transaction pool information. .

According to this, it is possible to find an attribute group consisting of an optimal combination of attributes for predicting the probability of rejection of an exchange deposit transaction, thereby improving the performance of predictability of rejection.

In relation to the genetic operator of selection evolution in step 4, first, in the 'selection' step, attribute groups having high predictability for rejection are stochastically selected. As a method of selection operation, Roulette wheel selection (the roulette area is allocated as much as it is proportional to the predicted degree of each attribute group, and then the roulette is selected to survive). Ranking selection (According to the predictive diagram, After ranking, select the attribute group that will survive with a predetermined probability), Elitist preserving selection (the most predictable attribute group is retained for the next generation), Tournament selection (optionally select multiple attribute groups, Among them, a method of selecting an attribute group having the highest predictability) can be used.

In the 'Crossover' stage, a new attribute group is created through mutual mating of attribute groups surviving in the generation. As a method of crossing operation, a simple crossover (one crossing position is set, and the genotype of the parent attribute groups are exchanged based on the front and rear), multi-point crossover, and crossover location Is plural), a uniform crossover (a method of filtering attributes that are genotype using a specific mask), and the like.

In the 'mutation' stage, after the recombination of the genotyping attribute is completed through mating, the attribute group is changed by replacing a specific attribute within the attribute group. As a method of variation calculation, a method of generating a random number for each attribute that is a genotype and mutating when the probability is lower than a specific variation probability may be used, and a static variation method that fixes the specific variation probability and a dynamic that dynamically adjusts the specific variation probability Mutation methods can be used. In particular, the dynamic variation method may be configured to increase the intensity of variation by increasing the probability of a specific variation in the initial generation, and to improve the quality by decreasing the probability of a specific variation as the generation progresses or after the quality of the household is improved by a certain level or higher. . According to the dynamic variation equation according to an embodiment of the present invention, the effect that the predictability can be improved quickly is generated in a generation having poor initial quality. According to this mutation step, an effect of searching globally even for combinations other than the initial gene combination is generated, thereby preventing the local minimum.

With respect to the prediction model based on reinforcement learning of the rejection probability prediction module 112, the rejection prediction prediction module 112 may be updated to the next step to generate rejectability information with improved accuracy.

(1) Attributes that make up the transaction information of the transaction (number of transactions sent from the account, fee information, amount of cryptocurrency to be transferred, size of data), and Attributes that make up the block information (number of blocks, blocks of uncle blocks List number, fee total, block difficulty, proof constants such as nonce and mix hash), and attributes that make up the transaction pool information (number of transactions included in the transaction pool, fee rank for a specific transaction, and the sum of fees in the transaction pool) The environment (Environment), the rejectability prediction module 112 as an agent (Agent), the rejection probability prediction module 112 for the transaction is determined whether the rejection action (Action), the actual rejection of the transaction Create a reinforcement learning model that rewards (Reward)

(2) The rejection probability prediction module 112, which is an agent, determines whether or not to reject the past transaction (predicts rejection probability) and performs an action

(3) Reward is performed to the rejection probability prediction module 112, which is an agent, based on whether or not the actual transaction has been rejected in the past.

(4) The rejection probability prediction module 112 is updated based on the reward.

(5) Repeat the above steps

According to an embodiment of the present invention, after a specific closing condition is satisfied, the final updated rejection probability prediction module 112 outputs rejection possibility information (or rejection determination) of the newly input exchange deposit transaction, and the deposit approval module It may be configured to transmit the rejectability information to 115.

According to this, an effect capable of improving the accuracy of the rejection determination (rejection probability prediction), which is an action of the rejection probability prediction module 112, without learning the uniform weight for the attribute is generated.

According to an embodiment of the present invention, the deposit approval module 115 may approve the deposit based on the rejectability information or rejection determination generated by the rejectability predictive module 112 to enable transactions within an exchange. have.

Alternatively, the dynamic fee calculation module 114 calculates the dynamic fee based on the rejection probability information generated by the rejection probability prediction module 112 or the rejection determination and confirmation time prediction information of the confirmation time prediction module 113, When the dynamic fee calculated by the user client 2 is approved, the deposit approval module 115 may approve the deposit for the corresponding cryptocurrency, so that the transaction can be configured within the exchange.

The confirmation time prediction module 113 is a property constituting the transaction information of the exchange deposit transaction (the number of transactions sent from the account, the fee information, the amount of cryptocurrency to be transferred, the size of data), and the property constituting the block information (block Number, block list number of uncle blocks, sum of fees, difficulty of blocks, proof constants such as nonce and mix hash), attributes that make up the transaction pool information (number of transactions in the transaction pool, commission rank of a specific transaction, The module generates prediction time prediction information by predicting the time when the exchange deposit transaction is recorded in the new block and confirmed by the block proof of the mining node based on the sum of the transaction pool fees). The confirmation time prediction for the exchange deposit transaction of the confirmation time prediction module 113 may be performed by an artificial intelligence-based prediction model such as basic machine learning, artificial neural network, genetic algorithm, and SVM using supervised learning or unsupervised learning. .

The dynamic fee calculation module 114 includes the rejection probability information generated by the rejection probability prediction module 112, the confirmation time prediction information predicted by the confirmation time prediction module 113, and transaction information of the corresponding cryptocurrency (transaction volume, unit price per hour, hourly (A), it is dynamically calculated on the basis of at least one of the fees applied to the immediate deposit authorization can be calculated and generated immediately. The instant deposit fee according to an embodiment of the present invention may be calculated based on the expected loss of the exchange that may occur when the deposit approval module 115 immediately approves the deposit for a specific exchange deposit transaction and is traded on the exchange. have. For example, the probability that an exchange deposit transaction will be rejected according to the rejection possibility information is '5%', and the time at which the block in which the exchange deposit transaction is recorded is proved by the miner is expected to be confirmed according to the confirmation time confirmation information. If it is '10 minutes ', and the amount of cryptocurrency requested by the user client 2 is' 100 ETH ', the dynamic fee calculation module 114 is based on the expected value '95 ETH' calculated based on the rejectability information. The maximum expected loss can be calculated as '14 .5 ETH 'by applying the variability calculated based on the confirmation time prediction information, and the dynamic fee calculation module 114 immediately converts the calculated maximum expected loss into a user deposit fee ( 2) can be configured to output. When the user client 2 immediately approves the deposit fee '14 .5ETH 'and transmits the fee approval information to the deposit approval module 115, the deposit approval module 115 is' 100 ETH, the quantity of cryptocurrency requested to be deposited. It can be configured to approve the deposit of '85 .5 ETH ', which is a limited quantity, of '14 .5 ETH' of the deposit fee immediately. Users will be able to freely trade '85 .5 ETH 'approved for deposit within the exchange immediately.

The deposit approval module 115 approves and approves the deposit to the user account of the user client 2 for a specific quantity of a specific cryptocurrency based on the possibility of rejection of the exchange deposit transaction or the instant deposit fee and fee approval information. It is a module that allows you to freely trade instantly on the amount of cryptocurrencies in the exchange.

When the deposit approval module 115 approves the deposit to the user account for a specific quantity of a specific cryptocurrency with only the rejection possibility information for the exchange deposit transaction, whether to approve using the threshold or cut off value of the rejection possibility information Can decide.

When the deposit approval module 115 immediately approves the deposit of the user account for a specific quantity of a specific cryptocurrency based on the deposit fee and the fee approval information, the fee approval information including the signature of the user account is received. Only when the payment is approved, the amount of the deposit fee immediately in the total quantity of a specific cryptocurrency can be set to enable immediate trading (sold) within the exchange only for a limited quantity.

Accordingly, if the deposit for the exchange deposit transaction is approved by the deposit approval module, before the proof by the mining node of the block in which the exchange deposit transaction is recorded, the user can perform the verification of the approved specific cryptocurrency. The effect is that it becomes possible to sell within the exchange. In other words, after users send a certain amount of cryptocurrency to the exchange account, it is possible to immediately trade (sell) within the exchange without waiting for a long time to confirm.

As described above, those skilled in the art to which the present invention pertains will understand that the present invention may be implemented in other specific forms without changing its technical spirit or essential features. Therefore, it should be understood that the above-described embodiments are illustrative in all respects and not restrictive. The scope of the present invention is indicated by the following claims rather than the detailed description, and all changes or modified forms derived from the meaning and scope of the claims and equivalent concepts should be interpreted to be included in the scope of the present invention.

The features and advantages described in this specification are not all inclusive, and many additional features and advantages will become apparent to those skilled in the art in view of the drawings, specifications, and claims. Moreover, it should be noted that the language used herein has been chosen primarily for readability and for teaching purposes, and may not be selected to describe or limit the subject matter of the present invention.

The foregoing description of embodiments of the invention has been presented for purposes of illustration. It is not intended to be exhaustive or to limit the invention to the precise form disclosed. Those skilled in the art can understand that many modifications and variations are possible in light of the above disclosure.

Therefore, the scope of the present invention is not limited by the detailed description, but by any claims of the application based thereon. Accordingly, the disclosure of the embodiments of the present invention is exemplary and does not limit the scope of the present invention as set forth in the claims below.

1: Cryptocurrency exchange deposit approval device
2: User client
3: Node
4: Connection node
10: Apparatus for approving deposits of cryptocurrency exchanges using artificial intelligence
30: Blockchain
100: node information collection module
101: node evaluation information generation module
102: node priority update module
103: Tx information collection module
104: Tx confirmation module
105: deposit approval module
110: transaction information collection module
111: block information collection module
112: rejection probability prediction module
113: confirmation time prediction module
114: dynamic fee calculation module
115: deposit approval module

Claims (2)

The decentralized network of the specific blockchain is composed of transaction information of an exchange deposit transaction, which is a transaction that transfers a specific cryptocurrency from a user account that is a user's account in a specific blockchain to an exchange account that is an account of an exchange in the specific blockchain. A transaction information collection module received from a connection node that is at least one node among a plurality of nodes;
A rejection probability prediction module including an artificial intelligence program code for predicting rejectability of the exchange deposit transaction based on the transaction information of the exchange deposit transaction and generating rejection probability information;
A dynamic fee calculation module that generates an instant deposit fee that is an expected loss dynamically calculated based on the rejection probability information generated by the rejection probability prediction module; And
A deposit approval module configured to approve the deposit for the exchange deposit transaction upon receiving approval of the user account for the immediate deposit fee;
Including,
The AI program code of the rejection probability prediction module,
Attributes constituting transaction information of a specific transaction (number of transactions sent from the user account, fee information set by the user account in the exchange deposit transaction, amount of the specific cryptocurrency to be transferred, and size of data), block information Attributes to configure (number of blocks, number of block lists of uncle blocks, sum of fees, difficulty and proof constants of blocks) and attributes to configure transaction pool information, which is the information of transaction pool, which is a collection of transactions stored in memory before being written to the block. (The number of transactions included in the transaction pool, the commission rank of the exchange deposit transaction, and the sum of the fees in the transaction pool) as an environment, the rejectability prediction module as an agent, and the specific transaction A reinforcement learning model generation step of generating an reinforcement learning model that outputs the rejectability information of the rejectability predictive module for an action as an action and rewards whether or not the specific transaction is actually rejected;
An action step in which the rejection probability prediction module, which is the agent, predicts a rejection possibility for a transaction in the past and outputs the rejectability information;
A reward step of performing the reward on the rejection probability prediction module, which is the agent, based on whether the actual transaction was rejected in the past; And
An updating step of updating the rejection probability prediction module based on the reward;
It is performed on the computer, and the rejection prediction module is updated by repeating the action step or the update step until a specific termination condition is satisfied,
The rejection probability prediction module generates the rejection probability information of the exchange deposit transaction after the specific closing condition is satisfied,
When the deposit for the exchange deposit transaction is approved by the deposit approval module, before the proof by the mining node of the current block, which is the block in which the exchange deposit transaction is recorded, is performed, the specific cryptocurrency that the user wishes to transfer Characterized in that configured to be possible to sell within the exchange amount of the immediate deposit fee from the amount of money,
Apparatus for instant deposit approval of cryptocurrency exchanges using reinforcement learning.
The transaction information collection module transmits a specific cryptocurrency from a user account, which is an account in a user's specific blockchain, to an exchange account, which is an account in the specific blockchain of the exchange, and the transaction information of the exchange deposit transaction, which is a transaction that transfers a specific cryptocurrency. Collecting transaction information received from a connection node that is at least one node among a plurality of nodes constituting a distributed network of the chain;
The rejection probability prediction module including an artificial intelligence program code for predicting the rejection probability of the exchange deposit transaction based on the transaction information of the exchange deposit transaction to generate rejection probability information, performs the artificial intelligence program code Predicting the possibility of rejection;
A dynamic fee calculation module, wherein the dynamic fee calculation module generates an instant deposit fee that is an expected loss dynamically calculated based on the rejection probability information generated by the rejection probability prediction module; And
A deposit approval step of, when the deposit approval module receives approval of the user account for the immediately deposit fee, depositing the deposit for the exchange transaction;
Including,
The AI program code of the rejection probability prediction module,
Attributes constituting transaction information of a specific transaction (number of transactions sent from the user account, fee information set by the user account in the exchange deposit transaction, amount of the specific cryptocurrency to be transferred, and size of data), block information Attributes to configure (number of blocks, number of block lists of uncle blocks, sum of fees, difficulty and proof constants of blocks) and attributes to configure transaction pool information, which is the information of transaction pool, which is a collection of transactions stored in memory before being written to the block. (The number of transactions included in the transaction pool, the commission rank of the exchange deposit transaction, and the sum of the fees in the transaction pool) as an environment, the rejectability prediction module as an agent, and the specific transaction A reinforcement learning model generation step of generating an reinforcement learning model that outputs the rejectability information of the rejectability predictive module for an action as an action and rewards whether or not the specific transaction is actually rejected;
An action step in which the rejection probability prediction module, which is the agent, predicts a rejection possibility for a transaction in the past and outputs the rejectability information;
A reward step of performing the reward on the rejection probability prediction module, which is the agent, based on whether the actual transaction was rejected in the past; And
An updating step of updating the rejection probability prediction module based on the reward;
It is performed on the computer, and the rejection prediction module is updated by repeating the action step or the update step until a specific termination condition is satisfied,
The rejection probability prediction module generates the rejection probability information of the exchange deposit transaction after the specific closing condition is satisfied,
When the deposit for the exchange deposit transaction is approved by the deposit approval module, before the proof by the mining node of the current block, which is the block in which the exchange deposit transaction is recorded, is performed, the specific cryptocurrency that the user wishes to transfer Characterized in that configured to be possible to sell within the exchange amount of the immediate deposit fee from the amount of money,
A method for promptly approving deposits at cryptocurrency exchanges using reinforcement learning.

Images