Artificial intelligence based used car overseas trading platform

Abstract

A method for providing an artificial intelligence-based used car overseas trading platform executed by a server processor, comprising: a step of translating and transmitting an inquiry text received from a buyer into the seller's language using a pre-trained artificial intelligence-based translation model; and a step of translating and transmitting a response text to the inquiry text received from the seller into the buyer's language using a translation model.

Description

Technology Field

[0001] The present disclosure relates to a platform, and more specifically, to an artificial intelligence-based overseas used car trading platform. Background Technology

[0002] With the development of internet information and communication and network technologies, interest in and utilization of e-commerce via the internet are rapidly expanding. Riding this trend, websites for buying and selling used cars online are being established one after another. For example, various online used car trading systems have been disclosed that receive used car information from sellers, database it, display it on a web page, determine prices, and are equipped with features that allow buyers to easily search for this sales information and support online transactions.

[0003] While such online used car trading systems have primarily been implemented through web applications, recently, used car trading systems utilizing mobile environments such as mobile communication devices are also being developed.

[0004] However, since existing used car trading systems do not provide translation into different languages, there is a problem where transactions between domestic export dealers and overseas import dealers cannot proceed smoothly. Prior art literature

[0005] Korean Published Patent No. 10-2018-0009625 The problem to be solved

[0006] The present disclosure aims to provide an artificial intelligence-based overseas used car trading platform.

[0007] The technical problems of the present disclosure are not limited to those mentioned above, and other unmentioned technical problems will be clearly understood by those skilled in the art from the description below. means of solving the problem

[0008] A method for providing an artificial intelligence-based used car overseas trading platform performed by a processor of a server according to an embodiment of the present disclosure comprises: receiving an inquiry text in the first language from a terminal of a buyer using the first language, inquiring about the description, price, and delivery date of the first used car; obtaining an inquiry text in the second language, which is the second language translated by using a pre-trained artificial intelligence-based first translation model—the second language being the language of the first sellers that is pre-stored—; transmitting the inquiry text in the second language to each of the terminals of the first sellers; requesting first response texts in the second language, which include the description, price, and delivery date of the first used car, from the terminals of the first sellers; receiving the first response texts in the second language from each of the terminals of the first sellers; and obtaining first translated response texts in the first language, which are the first response texts in the second language translated by using the first translation model. and may include the step of transmitting the first translation response texts of the first language to the buyer's terminal.

[0009] Additionally, if the first response texts in the second language are not received from the first sellers within a predetermined first period after the step of requesting the first response texts in the second language, the method may further include the step of obtaining the inquiry text in the third language, which is obtained by translating the inquiry text in the first language into the third language using a pre-trained AI-based second translation model—the third language being the language of the second sellers that is pre-stored—; the step of transmitting the inquiry text in the third language to each of the second sellers' terminals; the step of receiving the response texts in the third language from the second sellers' terminals, which include a description of the first used car, a price, and a delivery date; the step of obtaining the second translated response texts in the first language, which are obtained by translating each of the response texts in the third language into the first language using the second translation model; and the step of transmitting the second translated response texts in the first language to the buyer's terminal.

[0010] Additionally, if a purchase decision message for the first used car corresponding to the first-1 translation answer text of the first language among the first translation answer texts of the first language is not received from the terminal of the buyer within a predetermined second period, the method may further include the step of transmitting to the terminals of the first sellers, respectively, the lowest price among the prices of the first used car included in the first answer texts of the second language and the earliest possible delivery date among the possible delivery dates of the first used car; the step of requesting second answer texts of the second language from each of the terminals of the first sellers, which include a price lower than the lowest price and a date earlier than the earliest possible delivery date; the step of receiving the second answer texts of the second language from each of the terminals of the first sellers; the step of obtaining third translation answer texts of the first language, which are obtained by translating the second answer texts of the second language into the first language using the first translation model; and the step of transmitting the third translation answer texts of the first language to the terminal of the buyer.

[0011] Additionally, after the step of transmitting the inquiry text in the second language to each of the terminals of the first sellers, if the first response texts in the second language are not received from the terminals of the first sellers within a predetermined period, the method may further include the step of requesting the terminals of the first sellers for third response texts in the second language including a description, price, and delivery date of the second used car corresponding to the car model and year of the first used car; the step of receiving the third response texts in the second language from each of the terminals of the first sellers; the step of obtaining fourth translated response texts in the first language by translating the third response texts in the second language into the first language using the first translation model; and the step of transmitting the fourth translated response texts in the first language to the terminal of the buyer.

[0012] Additionally, the method may further include the step of requesting a response regarding satisfaction of the first translation response texts of the first language from the terminal of the buyer; the step of determining the score of the first sellers based on the score of satisfaction when the response regarding satisfaction is received from the terminal of the buyer; and the step of excluding the first sellers from the transmission of the new inquiry text when the score of the first sellers is less than a predetermined threshold value.

[0013] A server providing an AI-based used car overseas trading platform according to an embodiment of the present disclosure comprises: a processor; a memory; and a network unit; wherein the processor comprises: an operation of receiving an inquiry text in the first language inquiring about a description, price, and delivery date of a first used car from a terminal of a buyer using the first language; an operation of obtaining an inquiry text in the second language, which is translated into the first language using a pre-trained AI-based first translation model—the second language being the language of a first seller stored in advance—; an operation of transmitting the inquiry text in the second language to each of the terminals of the first sellers; an operation of requesting first response texts in the second language, which include a description, price, and delivery date of the first used car, from the terminals of the first sellers; an operation of receiving first response texts in the second language from each of the terminals of the first sellers; and an operation of obtaining first translated response texts in the first language, which are each of the first response texts in the second language translated into the first language using the first translation model. and the operation of transmitting the first translation response texts of the first language to the buyer's terminal can be performed.

[0014] Additionally, the processor may further perform the following operations: if, after the operation of requesting the first answer texts of the second language, the first answer texts of the second language are not received from the first sellers within a predetermined first period, the operation of obtaining the inquiry text of the third language, which is obtained by translating the inquiry text of the first language into the third language using a pre-trained AI-based second translation model—the third language being the language of the second sellers that is pre-stored—; the operation of transmitting the inquiry text of the third language to each of the second sellers' terminals; the operation of receiving the answer texts of the third language, which include the description, price, and delivery date of the first used car, from the terminals of the second sellers; the operation of obtaining the second translated answer texts of the first language, which are obtained by translating each of the answer texts of the third language into the first language using the second translation model; and the operation of transmitting the second translated answer texts of the first language to the buyer's terminal.

[0015] Additionally, if the processor fails to receive a purchase decision message for the first used car corresponding to the first-1 translation answer text of the first language among the first translation answer texts of the first language from the buyer's terminal within a predetermined second period, the processor may further perform the operation of transmitting to the terminals of the first sellers, respectively, the lowest price among the prices of the first used car included in the first answer texts of the second language and the earliest possible delivery date among the possible delivery dates of the first used car; the operation of requesting second answer texts of the second language from each of the terminals of the first sellers, which include a price lower than the lowest price and a date earlier than the earliest possible delivery date; the operation of receiving second answer texts of the second language from each of the terminals of the first sellers; the operation of obtaining third translation answer texts of the first language, which are obtained by translating each of the second answer texts of the second language into the first language using the first translation model; and the operation of transmitting the third translation answer texts of the first language to the buyer's terminal.

[0016] Additionally, the processor may further perform the following operations: after transmitting the inquiry text of the second language to each of the terminals of the first sellers, if the first answer texts of the second language are not received from the terminals of the first sellers within a predetermined period; requesting the terminals of the first sellers for third answer texts of the second language including a description, price, and delivery date of the second used car corresponding to the model and year of the first used car; receiving the third answer texts of the second language from each of the terminals of the first sellers; obtaining fourth translated answer texts of the first language by translating the third answer texts of the second language into the first language using the first translation model; and transmitting the fourth translated answer texts of the first language to the terminal of the buyer.

[0017] Additionally, the processor may further perform the operation of requesting a response regarding the satisfaction of the first translation response texts of the first language from the terminal of the buyer; the operation of determining the score of the first sellers based on the score of the satisfaction when the response regarding the satisfaction is received from the terminal of the buyer; and the operation of excluding the first sellers from the transmission of the new inquiry text when the score of the first sellers is less than a predetermined threshold value. Effects of the invention

[0018] The present disclosure enables the trading of used cars regardless of language by utilizing an artificial intelligence-based translation model.

[0019] The effects obtainable from the present disclosure are not limited to those mentioned above, and other unmentioned effects will be clearly understood by those skilled in the art to which the present disclosure belongs from the description below. Brief explanation of the drawing

[0020] FIG. 1 is a diagram showing a system for providing an artificial intelligence-based used car overseas trading platform according to one embodiment of the present disclosure. FIGS. 2 to 6 are flowcharts illustrating an example of a method for providing an artificial intelligence-based used car overseas trading platform performed by a processor of a server according to an embodiment of the present disclosure. FIGS. 7 and 8 are drawings for illustrating a first translation model according to one embodiment of the present disclosure. FIGS. 9 and 10 are drawings for illustrating a second translation model according to one embodiment of the present disclosure. Specific details for implementing the invention

[0021] Various embodiments are now described with reference to the drawings. In this specification, various descriptions are provided to provide an understanding of the present disclosure. However, it is evident that these embodiments can be practiced without such specific descriptions.

[0022] As used herein, terms such as "examples," "examples," "aspects," "examples," etc., may not be interpreted as implying that any aspect or design described is better or more advantageous than other aspects or designs.

[0023] The term "or" is intended to mean an intensional "or" rather than an exclusive "or." That is, unless otherwise specified or clear from the context, it is intended to mean one of the natural intensional substitutions of "X uses A or B."

[0024] Additionally, the terms “comprising” and / or “comprising” should be understood to mean that such features and / or components are present. However, the terms “comprising” and / or “comprising” should be understood not to exclude the presence or addition of one or more other features, components and / or groups thereof. Furthermore, unless otherwise specified or clearly evident from the context to indicate a singular form, the singular in this specification and claims should generally be interpreted to mean “one or more.”

[0025] FIG. 1 is a diagram showing a system for providing an artificial intelligence-based used car overseas trading platform according to one embodiment of the present disclosure.

[0026] Referring to FIG. 1, a system for providing an artificial intelligence-based used car overseas trading platform according to one embodiment of the present disclosure may include a buyer terminal (10), a seller terminal (20), a network (30), a server (100), etc.

[0027] In the present disclosure, an artificial intelligence-based overseas used car trading platform may refer to a device or system that receives requests related to the purchase and sale of used cars from users who access the platform for the primary purpose of trading used cars regardless of nationality and language, and performs operations related to the purchase and sale of used cars. In the present disclosure, the artificial intelligence-based overseas used car trading platform may be provided by a server (100).

[0028] A buyer terminal (10) may be a terminal of a buyer who wishes to purchase a used car. For example, a buyer terminal (10) may be a terminal of an overseas or domestic buyer who wishes to purchase a used car. There may be one or more buyer terminals (10). If there are multiple buyer terminals (10), the server (100) may group the buyer terminals (10) by language and store them in advance. A buyer terminal (10) may access an AI-based overseas used car trading platform provided by the server (100) and request the server (100) to purchase a used car.

[0029] The seller terminal (20) may be a terminal of a seller who intends to sell a used car. For example, the seller terminal (20) may be a terminal of an overseas or domestic seller who intends to sell a used car. The seller terminal (20) may have one or more units. If there are multiple seller terminals (20), the server (100) may group the seller terminals (20) by language and store them in advance. The seller terminal (20) may access the AI-based overseas used car trading platform provided by the server (100) and request the server (100) to sell the used car.

[0030] The network (30) may include any wired or wireless communication network capable of transmitting and receiving data and signals of any form. Specifically, the network (30) may include a wired or wireless communication network capable of transmitting and receiving data and signals between a buyer terminal (10), a seller terminal (20), and a server (100).

[0031] The server (100) may include one or more processors, storage units and communication units.

[0032] The processor may be composed of one or more cores. The processor may control the overall operation of the server (100). The processor may read a computer program stored in a storage unit to provide an artificial intelligence-based used car overseas trading platform according to one embodiment of the present invention.

[0033] The storage unit can store information of any form generated or determined by the processor and information of any form received through the communication unit. The storage unit may include at least one type of storage medium among flash memory type, hard disk type, multimedia card micro type, card type memory (SD, XD memory, etc.), RAM (Random Access Memory), SRAM (Static Random Access Memory), ROM (Read-Only Memory), EEPROM (Electrically Erasable Programmable Read-Only Memory), PROM (Programmable Read-Only Memory), magnetic memory, magnetic disk, and optical disk.

[0034] The communication unit may include any wired or wireless communication network capable of transmitting and receiving data, information, and signals of any form. The communication unit may communicate with an external device (e.g., buyer terminal (10), seller terminal (20)).

[0035] FIGS. 2 to 6 are flowcharts illustrating an example of a method for providing an artificial intelligence-based used car overseas trading platform performed by a processor of a server (100) according to one embodiment of the present disclosure.

[0036] The steps illustrated in FIGS. 2 through 6 are exemplary steps. Accordingly, it will also be apparent to those skilled in the art that some of the steps in FIGS. 2 through 6 may be omitted or additional steps may exist without departing from the scope of the present disclosure.

[0037] In one embodiment, the processor of the server (100) may provide the used car sales conditions of a seller using a second language to a terminal of a buyer using a first language. In one embodiment, the first language and the second language may be different languages.

[0038] For example, referring to FIG. 2, the processor of the server (100) can receive a text of inquiry in the first language from a terminal of a buyer using the first language, inquiring about the description, price, and delivery date of the first used car (210).

[0039] The description of the first used car may include the vehicle type, year of manufacture, color, accident location, insurance settlement amount, and documents related to accident verification.

[0040] The processor of the server (100) can obtain text of a text in a second language, which is translated into a second language, by using a pre-trained AI-based first translation model (220). A detailed description of the first translation model will be provided later with reference to FIGS. 7 and 8.

[0041] The second language may be the language of the first sellers that is pre-stored in the storage unit of the server (100). The storage unit of the server (100) may pre-store multiple sellers grouped by language.

[0042] The processor of the server (100) can transmit the text of the inquiry in the second language to each of the terminals of the first sellers (230).

[0043] The processor of the server (100) may request first response texts in a second language, including a description of the first used car, a price, and a delivery date, from the terminals of the first sellers (240).

[0044] The response texts are answers to the inquiry text and may be information provided for each item listed in the inquiry text.

[0045] The processor of the server (100) can receive first response texts in a second language from the terminals of the first sellers (250).

[0046] The processor of the server (100) can obtain first translation answer texts of the first language, each of which is translated into the first language using the first translation model (260).

[0047] The processor of the server (100) can transmit the first translation response texts of the first language to the buyer's terminal (270).

[0049] In one embodiment, if the processor of the server (100) does not receive a response text from the first sellers using the second language, it may translate the buyer's inquiry text into the third language and provide it to the second sellers using the third language, and request a response from the second sellers.

[0050] For example, referring to FIG. 3, if the processor of the server (100) fails to receive the first answer texts in the second language from the first sellers within a predetermined first period (e.g., within one hour, etc.) after the step of requesting the first answer texts in the second language, it may obtain the text of the inquiry in the third language, which is translated from the text of the inquiry in the first language into the third language, using a pre-trained AI-based second translation model (310). In one embodiment, the first language and the third language may be different languages. Also, the second language and the third language may be different languages.

[0051] A detailed description of the second translation model will be provided later with reference to Figures 9 and 10.

[0052] The third language may be the pre-stored language of the second sellers.

[0053] The processor of the server (100) can transmit the text of the inquiry in the third language to the terminals of the second sellers (320).

[0054] The processor of the server (100) can receive response texts in a third language from the terminals of the second sellers, including a description of the first used car, the price, and the delivery date (330).

[0055] The processor of the server (100) can obtain second translation answer texts of the first language, each of which is translated into the first language using a second translation model (340).

[0056] The processor of the server (100) can transmit the second translation response texts of the first language to the buyer's terminal (350).

[0058] In one embodiment, if the processor of the server (100) does not receive a purchase decision message from the buyer, it may request better conditions from sellers to offer better conditions to the buyer.

[0059] For example, referring to FIG. 4, if the processor of the server (100) does not receive a purchase decision message for a first used car corresponding to the first-1 translation answer text of the first language among the first translation answer texts of the first language from the buyer's terminal within a predetermined second period (e.g., within one day, etc.), it may transmit the lowest price among the prices of the first used car included in the first answer texts of the second language and the earliest delivery date among the delivery dates of the first used car to the terminals of the first sellers, respectively (410).

[0060] The processor of the server (100) can request second response texts in a second language from each of the terminals of the first sellers, including a price lower than the lowest price and a date earlier than the earliest possible delivery date (420).

[0061] The processor of the server (100) can receive second response texts in a second language from the terminals of the first sellers (430).

[0062] The processor of the server (100) can obtain third translation answer texts of the first language, each of which is translated into the first language by using the first translation model (440).

[0063] The processor of the server (100) can transmit the third translation response texts of the first language to the buyer's terminal (450).

[0065] In one embodiment, if the processor of the server (100) does not receive a response text from the first sellers, it may request a response text for another used car similar to the first used car.

[0066] For example, referring to FIG. 5, if the processor of the server (100) fails to receive first response texts in the second language from the terminals of the first sellers within a predetermined period after the step (230) of transmitting inquiry texts in the second language to each of the terminals of the first sellers, it may request third response texts in the second language from the terminals of the first sellers, including a description, price, and delivery date of the second used car corresponding to the car model and year of the first used car (510).

[0067] The processor of the server (100) can receive third response texts in a second language from the terminals of the first sellers (520).

[0068] The processor of the server (100) can obtain fourth translation answer texts of the first language, each of which is translated into the first language by using the first translation model (530).

[0069] The processor of the server (100) can transmit the fourth translation response texts of the first language to the buyer's terminal (540).

[0071] In one embodiment, the processor of the server (100) may request satisfaction from the buyer regarding the translation response texts and determine the sellers' scores based on the satisfaction.

[0072] For example, referring to FIG. 6, the processor of the server (100) may request a response from the buyer's terminal regarding the satisfaction of the first translation response texts of the first language (610).

[0073] When the processor of the server (100) receives a response regarding satisfaction from the buyer's terminal, it can determine the score of the first sellers based on the score of satisfaction (620).

[0074] The response regarding satisfaction may include a satisfaction score. The satisfaction score may include individual scores based on items (e.g., satisfaction with the description of the used car, satisfaction with the price of the used car, satisfaction with the delivery date of the used car, satisfaction with the time of the response, etc.), and may be the sum or average of the individual scores.

[0075] In one embodiment, the satisfaction score may be calculated by applying a weight to each item. The item-specific weight may be determined based on the items included in the pre-stored responses regarding the buyers' satisfaction. In one embodiment, the item-specific weight may be assigned higher the more responses there are for each item included in the pre-stored responses regarding the buyers' satisfaction. For example, the pre-stored responses regarding the first buyer's satisfaction may only include satisfaction with the description of the used car and satisfaction with the price of the used car. And the pre-stored responses regarding the second buyer's satisfaction may only include satisfaction with the description of the used car. In this case, the weight for satisfaction with the description of the used car may be the highest. Next, the weight for satisfaction with the price of the used car may be the second highest. The weights for satisfaction with the delivery date of the used car and satisfaction with the time of the response may be the lowest.

[0076] In one embodiment, the satisfaction score can be calculated by the following mathematical formula 1.

[0077]

[0078] Here, x can be a satisfaction score. S is the satisfaction with the used car description, and w S may be the weight of satisfaction with the description of the used car. P is satisfaction with the price of the used car, and w P may be the weight of satisfaction regarding the price of the used car. D is satisfaction regarding the delivery date of the used car, and w D may be the weight of satisfaction regarding the available delivery date of the used car. T is satisfaction regarding the time of the response, and w T may be a weight of satisfaction regarding the time of the response. In one embodiment, the sum of the weights (w S + w P + w D + w T) can be 1. The ranges for satisfaction with the used car description (S), satisfaction with the used car price (P), satisfaction with the available delivery date (D), and satisfaction with the response time (T) can each be 0 to 10. Therefore, the maximum score for satisfaction can be 10 points. For example, the value of each weight is w S = 0.4, w P = 0.3, w D = 0.2, w T If = 0.1 and the individual scores for each item are S = 10, P = 9, D = 8, T = 7, the satisfaction score may be 5.4 points.

[0080] The processor of the server (100) may exclude the first sellers from the transmission of new inquiry texts if the score of the first sellers is below a predetermined threshold (e.g., 5 points in the case of a maximum of 10 points). Accordingly, the processor of the server (100) may improve the satisfaction of the future buyer by determining the sellers to be presented to the buyer in the future based on the buyer's satisfaction.

[0082] In one embodiment, when the processor of the server (100) receives a purchase decision message regarding a used car from a buyer, it may request better conditions from sellers other than the seller selling the used car. Accordingly, the processor of the server (100) can provide the buyer with the best conditions for the used car by inducing competition among sellers.

[0083] For example, when the processor of the server (100) receives a purchase decision message for a first used car corresponding to the first-2 translation answer text of the first language among the first translation answer texts of the first language from the buyer's terminal, it may transmit the first-2 translation answer text to the remaining terminals among the terminals of the first sellers, excluding the terminal of the first-1 seller that provided the first-2 translation answer text of the first language. In the present disclosure, the first-1 translation answer text of the first language and the first-2 translation answer text of the first language may be identical or different from each other.

[0084] The processor of the server (100) may request the remaining terminals to provide fourth answer texts in the second language that include a price lower than the price of the first used car and a date earlier than the delivery date included in the first-2 translation answer texts of the first language.

[0085] If the processor of the server (100) fails to receive the fourth response texts of the second language from the remaining terminals for a predetermined third period (e.g., one hour, etc.), it may transmit information about the first-1 seller (e.g., the first-1 seller's phone number, email, personal information, etc.) to the buyer.

[0086] When the processor of the server (100) receives the fourth answer texts of the second language from the remaining terminals, it can obtain the fourth translated answer texts of the first language by using the first translation model to translate the fourth answer texts of the second language into the first language.

[0087] The processor of the server (100) can transmit the fourth translation response texts of the first language to the buyer's terminal.

[0089] FIGS. 7 and 8 are drawings for illustrating a first translation model according to one embodiment of the present disclosure.

[0090] Referring to FIG. 7, the processor of the server (100) can obtain a text of a second language inquiry (730) by inputting a text of a first language inquiry (710) into a pre-trained AI-based first translation model (720).

[0091] Referring to FIG. 8, the processor of the server (100) can obtain a translation response text (830) of the first language by inputting a response text (810) of the second language into a pre-trained AI-based first translation model (720).

[0092] A first translation model can be pre-trained using training data that includes information about a first language and information about a second language to output text in a second language (or text in a first language) as a response when text in a first language (or text in a second language) is input.

[0093] FIGS. 9 and 10 are drawings for illustrating a second translation model according to one embodiment of the present disclosure.

[0094] Referring to FIG. 9, the processor of the server (100) can obtain a text of a question in a third language (930) by inputting a text of a question in a first language (910) into a pre-trained AI-based second translation model (920).

[0095] Referring to FIG. 10, the processor of the server (100) can obtain a translation response text (1030) of the first language by inputting a response text (1010) of the third language into a pre-trained AI-based second translation model (1020).

[0096] A second translation model can be pre-trained using training data that includes information about a first language and information about a third language, so that when text in the first language (or text in the third language) is input, it outputs text in the third language (or text in the first language) as a response.

[0097] An AI-based translation model (e.g., a first translation model, a second translation model, etc.) may be composed of a set of interconnected computational units that can generally be referred to as nodes. These nodes may also be referred to as neurons. An AI-based translation model may be composed of at least one node. The nodes (or neurons) constituting the AI-based translation model may be interconnected by one or more links.

[0098] In an AI-based translation model (e.g., a first translation model, a second translation model, etc.), one or more nodes connected via links can form a relative relationship between an input node and an output node. The concepts of input and output nodes are relative; any node in an output node relationship with respect to one node may be in an input node relationship with respect to another node, and vice versa. As described above, the relationship between an input node and an output node can be generated based on links. One or more output nodes may be connected to a single input node via links, and vice versa.

[0099] In a relationship between an input node and an output node connected through a single link, the value of the output node's data can be determined based on the data input into the input node. Here, the link interconnecting the input node and the output node may have a weight. The weight may be variable and may be varied by a user or an algorithm to enable an AI-based translation model (e.g., a first translation model, a second translation model, etc.) to perform the desired function. For example, if one or more input nodes are interconnected to a single output node by respective links, the output node's value can be determined based on the values ​​input into the input nodes connected to the output node and the weight set on the link corresponding to each input node.

[0100] An AI-based translation model (e.g., a first translation model, a second translation model, etc.) may have the same number of nodes in the input layer as the number of nodes in the output layer, and may be a neural network in which the number of nodes decreases and then increases again as it progresses from the input layer to the hidden layer. Additionally, a neural network according to another embodiment may have the same number of nodes in the input layer as the number of nodes in the output layer, and may be a neural network in which the number of nodes decreases as it progresses from the input layer to the hidden layer. Furthermore, an AI-based translation model according to yet another embodiment may have the same number of nodes in the input layer as the number of nodes in the output layer, and may be a neural network in which the number of nodes increases as it progresses from the input layer to the hidden layer. An AI-based translation model according to yet another embodiment may be a neural network in which the above-described neural networks are combined.

[0101] A deep neural network (DNN) included in an AI-based translation model (e.g., a first translation model, a second translation model, etc.) may refer to a neural network that includes multiple hidden layers in addition to an input layer and an output layer. Using a deep neural network allows for the identification of latent structures of data. A deep neural network may include a convolutional neural network (CNN), a recurrent neural network (RNN), an autoencoder, etc. The description of the deep neural network described above is merely an example and the present disclosure is not limited thereto.

[0102] AI-based translation models (e.g., first translation model, second translation model, etc.) can be trained through supervised learning. The training of an AI-based translation model may be a process of applying knowledge to the AI-based translation model to perform a specific action.

[0103] AI-based translation models (e.g., the first translation model, the second translation model, etc.) can be trained to minimize output errors. The training of an AI-based translation model may involve repeatedly inputting training data into the model, calculating the error between the model's output and the target for the training data, and updating the weights of each node of the model (e.g., the first translation model, the second translation model, etc.) by backpropagating the error from the output layer to the input layer in a direction that reduces the error. In the case of supervised learning, training data with correct answers labeled for each training data point can be used. For example, the training data may consist of data where each training data point is labeled with a category. The labeled training data is input into the AI-based translation model, and the error can be calculated by comparing the model's output with the labels of the training data.

[0104] It should be understood that the specific order or hierarchy of steps in the presented processes is an example of exemplary approaches. It should be understood that the specific order or hierarchy of steps in the processes may be rearranged within the scope of this disclosure based on design priorities. The appended method claims provide elements of various steps in a sample order, but do not imply being limited to the specific order or hierarchy presented.

[0105] Description of the presented embodiments is provided so that a person skilled in the art may use or practice the present disclosure. Various modifications to these embodiments will be apparent to a person skilled in the art, and the general principles defined herein may be applied to other embodiments without departing from the scope of the present disclosure. Thus, the present disclosure is not limited to the embodiments presented herein, but should be interpreted in the broadest possible scope consistent with the principles and novel features presented herein. Explanation of the symbols

[0106] 10: Buyer's terminal, 20: Seller's terminal, 30: Network, 100: Server

Claims

Claim 1 A method for providing an AI-based overseas used car trading platform executed by a server processor, comprising: receiving an inquiry text in the first language from a terminal of a buyer using the first language, inquiring about the description, price, and delivery date of the first used car; obtaining an inquiry text in the second language, which is obtained by translating the inquiry text in the first language into the second language using a pre-trained AI-based first translation model—the second language being the language of the first sellers stored in advance—; transmitting the inquiry text in the second language to each of the terminals of the first sellers; requesting first response texts in the second language, which include the description, price, and delivery date of the first used car, from the terminals of the first sellers; receiving the first response texts in the second language from each of the terminals of the first sellers; obtaining first translated response texts in the first language, which are obtained by translating each of the first response texts in the second language into the first language using the first translation model; transmitting the first translated response texts in the first language to the terminal of the buyer; and transmitting to the terminal of the buyer the The method comprises: a step of requesting a response regarding the satisfaction of first translation response texts of a first language; a step of determining the scores of the first sellers based on the satisfaction score when the response regarding the satisfaction is received from the buyer's terminal; and a step of excluding the first sellers from the transmission of new inquiry texts when the scores of the first sellers are below a predetermined threshold value; wherein the satisfaction score comprises S, satisfaction regarding the description of the used car; P, satisfaction regarding the price of the used car; D, satisfaction regarding the possible delivery date of the used car; T, satisfaction regarding the time of the response; and w, a weight corresponding to each. S , w P , w D , w T Based on, it is characterized by being calculated by the following mathematical formula 1, [Mathematical Formula 1] Here, x is the satisfaction score, and S, P, D, and T each have a range of 0 to 10, and w S , w P , w D , w T are each set to have a value greater than 0, and the above w S + w P + w D + w T is the denominator term of the above mathematical formula 1 (w) whether normalized to 1 or not. S + w P + w D + w T Characterized by being configured to be normalized by ), and said weight w S , w P , w D , w T ...is determined based on the frequency of responses by item included in the responses regarding the satisfaction of buyers pre-stored on the server, and is characterized by being configured such that the higher the response frequency for a specific item, the greater the weight assigned to that item; the threshold is set to 5 points on a scale of 10 points, and the scores of the first sellers are accumulated or updated for each first seller based on the satisfaction score x; and after the step of requesting the first response texts in the second language, if the first response texts in the second language are not received from the first sellers within a predetermined first period, the method comprises the step of obtaining the inquiry text in the third language, which is obtained by translating the inquiry text in the first language into the third language using a pre-trained AI-based second translation model—the third language is the language of the second sellers pre-stored—; the step of transmitting the inquiry text in the third language to each of the terminals of the second sellers; the step of receiving the response texts in the third language, including the description, price, and possible delivery date of the first used car, from the terminals of the second sellers; and using the second translation model, the The method further comprises the steps of: obtaining second translated answer texts of the first language, each of which is translated into the first language; transmitting the second translated answer texts of the first language to the buyer's terminal; and, if a purchase decision message for the first used car corresponding to the first-1 translated answer text of the first language among the first translated answer texts of the first language is not received from the buyer's terminal within a predetermined second period, transmitting the lowest price among the prices of the first used car included in the first answer texts of the second language and the earliest possible delivery date among the possible delivery dates of the first used car to the terminals of the first sellers, respectively.A step of requesting second response texts in a second language, including a price lower than the lowest price and a date earlier than the earliest possible delivery date, from each of the terminals of the first sellers; a step of receiving the second response texts in the second language from each of the terminals of the first sellers; a step of obtaining third translated response texts in the first language, each of which is translated into the first language using the first translation model. The method further comprises the step of transmitting the third translated answer texts of the first language to the buyer's terminal; and, after the step of transmitting the inquiry text of the second language to the terminals of the first sellers, if the first answer texts of the second language are not received from the terminals of the first sellers within a predetermined period, the step of requesting the third answer texts of the second language from the terminals of the first sellers, including a description, price, and delivery date of the second used car corresponding to the model and year of the first used car; the step of receiving the third answer texts of the second language from the terminals of the first sellers; the step of obtaining the fourth translated answer texts of the first language, which are obtained by translating the third answer texts of the second language into the first language using the first translation model; and the step of transmitting the fourth translated answer texts of the first language to the buyer's terminal.

Images