KR20210059253A - Method of providing a fulfillment service and service system therefor - Google Patents

Abstract

According to one embodiment of the present invention, provided is a method for providing a fulfillment service in a service server between an online seller and an offline warehouse, which comprises the steps of: registering service keywords through artificial intelligence learning; receiving order information for product shipment from an online seller terminal; extracting search data in the received order information and classifying the extracted search data; comparing the classified search data with registered service keywords; generating first fulfillment service data related to the corresponding service keyword according to the comparison result; and transmitting the generated first fulfillment service data to an offline warehouse terminal to control the product to be shipped out.

Description

TECHNICAL FIELD [Method OF PROVIDING A FULFILLMENT SERVICE AND SERVICE SYSTEM THEREFOR}

The present invention relates to a full-fillment service, and more particularly, to a method for providing a full-fillment service through automatic mapping of keywords to order information based on artificial intelligence, and a service system therefor.

In the conventional offline commerce, as media and consumption are gradually personalized, online commerce has been activated. Such online commerce is also moving from the one-person media era to the one-person customized consumption era, and now to the one-person seller era with the advancement of mobile environment and ICT technology.

In this process, the new generation, which has emerged as a core consumption group, gradually trusts each other as influencers, and searches, purchases, and shares products that satisfy their tastes. Due to such changes in the environment, in the new generation, individual sellers that manufacture or source products that more easily satisfy their tastes and sell them through online malls are increasing.

However, in the case of one-person sellers or individual sellers, there is a realistic problem in terms of logistics processing along with the establishment of online sales channels.

In addition, small and medium-sized warehouse owners are also faced with difficulties as there is a demand for change of constitution that requires frequent stocking and releasing of products in small quantities of various types.

Therefore, it is required to develop a system for online commerce that meets new trends between sellers and warehouse owners.

An object of the present invention is to provide a full-filment service through automatic mapping of keywords to order information based on 　 artificial intelligence.

Another object of the present invention is to provide a service platform for a full-filment service between an online seller and an offline warehouse.

Another object of the present invention is to provide a computing device for the above-described method and a full-filment service system including the same.

The technical problems to be achieved in the present invention are not limited to the above technical problems, and other technical problems not mentioned 　 will be clearly understood by those of ordinary skill in the technical field to which the present invention belongs from the following description.

A method of providing a full-filment service in a service server between an online seller and an offline warehouse according to an embodiment of the present invention for solving the above-described problem includes: registering a service keyword through artificial intelligence learning; Receiving order information for releasing goods from the online seller terminal; Extracting search data in the received order information and classifying the extracted search data; Comparing the classified search data with registered service keywords; Generating first full-filment service data related to a corresponding service keyword according to the comparison result; And transmitting the generated first full-filment service data to the offline warehouse terminal to control the goods to be shipped.

In a method for providing a full-filment service in a service server between an online seller and an offline warehouse according to an embodiment of the present invention, the received order information includes at least one of text data, image data, and audio data. It is done.

In a method for providing a full-filment service in a service server between an online seller and an offline warehouse according to an embodiment of the present invention, the generated first full-filment service data is transmitted to the offline warehouse terminal to control the delivery of the corresponding product. The step of performing is performed after the confirmation data including the generated first full-filment service data is transmitted to the online seller terminal and feedback data for the confirmation data is received from the online seller terminal.

In a method for providing a full-filment service in a service server between an online seller and an offline warehouse according to an embodiment of the present invention, at least one of the first full-filment service data and the confirmation data is the new service keyword and the service At least one of information on the goods received in the offline warehouse corresponding to the keyword, information on the quantity of goods orders extracted from the order information, inventory information on the corresponding goods, and total inventory information on the goods received by the online seller in the offline warehouse It characterized in that it comprises one or more.

A method of providing a full-filment service in a service server between an online seller and an offline warehouse according to another embodiment of the present invention includes: registering a service keyword through artificial intelligence learning; Receiving order information for releasing goods from the online seller terminal; Extracting search data in the received order information, classifying the extracted search data, and comparing the classified search data with registered service keywords; If there is no service keyword matching the comparison result, generating a new service keyword and generating first full-filment service data related to the generated new service keyword; And transmitting the generated first full-filment service data to the offline warehouse terminal to control the goods to be shipped.

In a method for providing a full-filment service in a service server between an online seller and an offline warehouse according to an embodiment of the present invention, the generated first full-filment service data is transmitted to the offline warehouse terminal to control the delivery of the corresponding product. The step of transmitting the confirmation data including the generated first full-filment service data to the online seller terminal, characterized in that it is performed after the feedback data for the confirmation data is received from the online seller terminal.

In a method for providing a full-filment service in a service server between an online seller and an offline warehouse according to an embodiment of the present invention, registering the generated new service keyword as a service keyword and updating a service keyword database; further comprising It characterized in that it is made by doing.

In a method for providing a full-filment service in a service server between an online seller and an offline warehouse according to an embodiment of the present invention, the registering of the new service keyword comprises: When the feedback data for the confirmation data is received from the online seller terminal by transmitting to the online seller terminal, it is performed based on the received feedback data.

A method for providing a full-filment service in a service server between an online seller and an offline warehouse according to an embodiment of the present invention, the method comprising: parsing the feedback data received from the online seller terminal; And identifying the parsed feedback data content.

In a method for providing a full-filment service in a service server between an online seller and an offline warehouse according to an embodiment of the present invention, if the content of the identified feedback data is a negative message, extraction/classification of search data in the received order information, and Comparing the classified search data with a registered service keyword, and if there is no service keyword matching the re-comparison result, a new service keyword is generated, and second full-filment service data related to the generated new service keyword is And generating second full-filment service data according to the matched service keyword, if there is a service keyword matching the re-comparison result.

A service server providing a full-filment service between an online seller and an offline warehouse according to an embodiment of the present invention includes: a communication interface unit for receiving order information for discharging goods from the online seller terminal; A data extraction/classification unit for extracting search data in the received order information and classifying the extracted search data; And registering a service keyword through artificial intelligence learning, comparing the classified search data with the registered service keyword, generating first full-filment service data related to the service keyword according to the comparison result, and generating the first full-filment service data related to the service keyword according to the comparison result. And a control unit that transmits full-filment service data to the offline warehouse terminal to control the delivery of the corresponding article.

In a service server providing a full-filment service between an online seller and an offline warehouse according to an embodiment of the present invention, the received order information includes at least one of text data, image data, and audio data. .

In a service server providing a full-filment service between an online seller and an offline warehouse according to an embodiment of the present invention, the control unit transmits confirmation data including the generated first full-filment service data to the online seller terminal. Thus, after the feedback data for the confirmation data is received from the online seller terminal, the generated first full-filment service data is transmitted to the offline warehouse terminal to control the delivery of the corresponding article.

In a service server providing a full-filment service between an online seller and an offline warehouse according to an embodiment of the present invention, at least one of the first full-filment service data and the confirmation data is included in the new service keyword and the service keyword. At least one or more of information on goods received in a corresponding offline warehouse, product order quantity information extracted from the order information, inventory information for the product, and total inventory information on goods received by the online seller in the offline warehouse It characterized in that it comprises a.

According to another embodiment of the present invention, a service server providing a full-filment service between an online seller and an offline warehouse includes: a communication interface unit for receiving order information for discharging goods from the online seller terminal; A data extraction/classification unit for extracting search data in the received order information and classifying the extracted search data; And registering a service keyword through artificial intelligence learning, comparing the classified search data with the registered service keyword, and if there is no service keyword matching the comparison result, a new service keyword is generated, and the generated new service keyword And a control unit that generates first full-filment service data related to and transmits the generated first full-filment service data to the offline warehouse terminal to control the delivery of a corresponding article.

In a service server providing a full-filment service between an online seller and an offline warehouse according to an embodiment of the present invention, the control unit transmits confirmation data including the generated first full-filment service data to the online seller terminal. Thus, after the feedback data for the confirmation data is received from the online seller terminal, the generated first full-filment service data is transmitted to the offline warehouse terminal to control the delivery of the corresponding article.

In a service server providing a full-filment service between an online seller and an offline warehouse according to an embodiment of the present invention, the control unit controls to register the generated new service keyword as a service keyword and update the service keyword database. It is characterized.

In a service server providing a full-filment service between an online seller and an offline warehouse according to an embodiment of the present invention, the control unit transmits confirmation data including the generated first full-filment service data to the online seller terminal. Thus, when feedback data for the confirmation data is received from the online seller terminal, the new service keyword is registered based on the received feedback data.

In a service server providing a full-filment service between an online seller and an offline warehouse according to an embodiment of the present invention, the control unit parses the feedback data received from the online seller terminal, and identifies the parsed feedback data content. Characterized in that.

In a service server providing a full-filment service between an online seller and an offline warehouse according to an embodiment of the present invention, the control unit extracts/ The classified and classified search data is re-compared with the registered service keyword, but if there is no service keyword matching the re-comparison result, a new service keyword is generated, and second full-filment service data related to the generated new service keyword is Control to generate or, if there is a service keyword matching the re-comparison result, control to generate second full-filment service data according to the matching service keyword.

According to the present invention as described above, the effects described below can be obtained. However, the effect that can be obtained through the present invention is not limited thereto.

First, there is an effect of providing a full-filment service through automatic mapping of keywords to order information based on artificial intelligence.

Second, there is an effect of providing a service platform for full-filment services between online sellers and offline warehouses.

Third, there is an effect of providing a computing device for the above method and a full-filment service system including the same.

1 is a diagram illustrating a service system for providing a full-filment service between an online seller and an offline warehouse according to an embodiment of the present invention.
2 is a diagram illustrating a function of the full-filment service server 120 according to an embodiment of the present invention.
3 is a block diagram illustrating a configuration of a full-filment service server 120 according to an embodiment of the present invention.
4 is a block diagram illustrating a configuration of a full-filment service server 120 according to another embodiment of the present invention.
5 to 7 are diagrams for explaining performing a full-filment service based on an image included in order information in the full-filment service server 120 according to the present invention.
8 is a flowchart illustrating the concept of a full-filment service.
9 is a diagram illustrating a full-filment service process in the full-filment service system 100 according to an embodiment of the present invention.
10 is a diagram illustrating a method of providing a full-filment service according to an embodiment of the present invention.
11 is a diagram illustrating a method of providing a full-filment service according to another embodiment of the present invention.
12 is a diagram illustrating a management system UI for a full-filment service according to another embodiment of the present invention.
13 is a diagram illustrating a method of providing a full-filment service according to another embodiment of the present invention.
14 is a diagram illustrating a method of providing a full-filment service according to another embodiment of the present invention.
15 is a diagram illustrating a method of providing a full-filment service according to another embodiment of the present invention.

Hereinafter, with reference to the accompanying drawings, preferred embodiments according to the present invention will be described in detail. The detailed description to be disclosed hereinafter together with the accompanying drawings is intended to describe exemplary embodiments of the present invention, and is not intended to represent the only embodiments in which the present invention may be practiced.

Only these embodiments are provided to make the posting of the present invention complete, 　 to fully inform the scope of the invention to those of ordinary skill in the art to which the present invention belongs, and 　 the present invention to be defined by the scope of the claims. Only.

In some cases, in order to avoid obscuring the concept of the present invention, well-known structures and devices may be omitted, or may be illustrated in block diagram form centering on the core functions of each structure and device. In addition, the same components will be described using the same reference numerals throughout this specification.

Throughout the specification, 　when a certain part 　"comprising 　 or 　including", 　 means that other components may be further included rather than excluding other components unless otherwise stated. do.

In addition, the term "?? unit" described in the specification refers to a unit that processes at least one function or operation, which may be implemented by hardware or software or a combination of hardware and software. Furthermore, "one (a 　 or 　an)", "one", 　 and similar related words are all in the singular and plural unless otherwise indicated in the specification or clearly contradicted by the context in the context describing the present invention. It can be used in a sense including.

In addition, 　specific terms used in the embodiments of the present invention are provided to aid understanding of the present invention, and 　unless defined otherwise, 　all terms used herein including technical or scientific terms are It has the same meaning as commonly understood by those of ordinary skill in the art to which it belongs. The use of these specific terms may be changed in other forms without departing from the technical spirit of the present invention.

Hereinafter, even though it is referred to as an online seller, an off-line warehouse, etc. in this specification, this does not mean a user, such as a person, but rather the online seller, offline warehouse, and computing to be described later. As a computing device, it may mean a means capable of communicating with a fulfillment service server, for example, a terminal of an online seller, a terminal of an offline warehouse, and the like.

Meanwhile, in the above, all of the terminals and devices may include software for driving them based on hardware. For example, the full-fillment service server may perform various operations based on a fulfillment service platform to be described later as a computing device. In addition, these devices are not limited to stationary devices such as TVs and PCs, and include mobile devices such as mobile phones, smartphones, tablet PCs, and notebook computers.

In addition, hereinafter, various configurations described in the present specification or illustrated by drawings are mainly described in the singular form to aid understanding of the present invention and for convenience of description, but the present invention is not necessarily limited thereto. Therefore, even if it is described in the singular through context or the like or shown in the singular form in the drawings, it may have a plural meaning.

Hereinafter, various embodiments according to the present invention will be described with reference to the accompanying drawings, as follows.

1 is a diagram showing a service system for providing a full-filment service between an online seller and an offline warehouse according to an embodiment of the present invention.

Referring to FIG. 1, a system for providing a full-fillment service (hereinafter referred to as a “fulfillment service server”) 100 includes an online seller 110 and an offline warehouse 130, A computing device 120 for providing a full-filment service is included between the online seller 110 and the offline warehouse 130.

Accordingly, both devices, that is, the online seller 110 and the offline warehouse 130 may be connected to each other through the full-filment service server 120.

The online seller 110 may perform data communication with the full-filment service server 120 through a wired/wireless communication network.

Here, the wired/wireless communication network supports a communication protocol so that various data can be exchanged between components of the full-filment service system 100 according to the present invention, and the type is not particularly limited. If necessary, it is possible to support this even when data is collected through the web or the like through the full-filment service server 120 or the like. According to one embodiment, 　the wired/wireless communication network is, 　Internet Protocol (IP) network supporting large-capacity data communication through Internet Protocol (IP), or All IP integrating different 　IP　 networks. I can. According to another embodiment, the wired/wireless communication network is a wired network, a mobile communication network including WCDMA, a high speed downlink packet access (HSDPA), and a mobile communication network including Long Term Evolution (LTE), LTE advanced (LTE-A). ), a mobile communication network, a satellite communication network, and a Wi-Fi network, or may be formed by a combination of at least two or more of them. Bluetooth communication can also be supported here.

The full-filment service server 120 is a component that supports and performs full-filment service provision through the offline warehouse 130 at the request of the online seller 110, and may be located remotely, but is limited thereto. no.

As described above, this full-filment service server 120 refers to hardware/software (Hardware/Software) that implements the full-filment service platform according to the present invention to generate, process, and store various data, or support it. As such, depending on the embodiment, it may be described as being named as a processor, a controller, and the like, but it should be noted in advance that it should not be limited and interpreted by the name.

On the other hand, according to an embodiment, 　 full-filment service server 120, 　 Cloud 　 server, IMS (IP Multimedia Subsystem) 　 server, 　 Telephony Application (Telephony Application) 　 server, IM (Instant Messaging) 　 Server, MGCF (Media Gateway Control) Function) 　 server, MSG (Messaging Gateway) 　 server, CSCF (Call Session Control Function) 　 can be implemented in any one of the form of server.

Meanwhile, the full-filment service server 120 may provide a file for installing a full-filment service or a related application type program or software to an online seller 110 connected through a wired/wireless communication network. In this case, the online seller 110 may download or/and install an application using a file provided from the full-filment service server 120. In addition, the online seller 110 can receive or use the full-filment service by being connected to the full-filment service server 120 through its operating system or program.

In particular, the full-filment service server 120 receives order information from the online seller 110 and compares it with keyword data learned through artificial intelligence from the received order information according to the present invention. A full-filment service may be provided to the online seller 110 by searching for.

Meanwhile, although not shown, in FIG. 1, the full-filment service server 120 may be implemented as an internal component of the online seller 110 in the form of a single module, or any of the internal components of the online seller 110 One may perform some or all of the functions of the full-filment service server 120. For example, when the full-filment service according to the present invention is performed through an application installed in the online seller 110, an internal control configuration such as a central processing unit (CPU) of the online seller 110 is used for the full-filment service. It may replace or partially perform the functions of the server 120.

2 is a diagram illustrating a function of a full-filment service server 120 according to an embodiment of the present invention, and FIG. 3 is a diagram illustrating a configuration of a full-filment service server 120 according to an embodiment of the present invention. Figure 4 is a block diagram for explaining the configuration of the full-filment service server 120 according to another embodiment of the present invention, Figures 5 to 7 are in the full-filment service server 120 according to the present invention A diagram illustrating the execution of a full-filment service based on an image included in order information, FIG. 8 is a flow chart illustrating the concept of a full-filment service, and FIG. 9 is a full-filment service according to an embodiment of the present invention. A diagram illustrating a full-filment service process in the system 100, and FIGS. 10 to 11 are diagrams illustrating a method for providing a full-filment service according to the present invention, and FIGS. 12 to 15 are A diagram illustrating a method of providing various full-filment services according to an embodiment.

Hereinafter, various methods of providing full-filment services will be described to aid understanding of the present invention and for convenience of description. In this case, although the steps are sequentially described for the sake of explanation in the accompanying drawings or description, they are not necessarily limited to the illustrated order. For example, some of the steps may be performed simultaneously or in different steps than the order shown.

2, the full-filment service server 120 according to an embodiment of the present invention may participate in or support the following functions.

That is, the full-filment service server 120 may be involved in or process product warehousing, storage, order, packaging, delivery, return/inventory management, settlement, distribution support, and the like.

In connection, referring to FIG. 8, the concept of a full-filment service will be briefly described as follows.

For example, when a consumer in a first country (US) places an order to a seller in a second country (AU) through online (S110), the second seller finds the product in an offline warehouse after receiving the order, packs it, and then sends it. Request (S120).

Accordingly, the goods are shipped to the first country after declaration to the customs (S130), and finally delivered last mile to consumers in the first country (S140).

This whole process can be called full-filment service.

Referring to FIG. 3, a block diagram of a full-filment service server 120 according to an embodiment of the present invention is shown. Hereinafter, while describing FIGS. 13 and 14, the configuration of FIG. 3 will be described together.

Basically, the full-filment service server 120 includes a communication interface unit 310, a data extraction/classification unit 320, a data processing unit 330, a control unit 340, and a database 350.

According to an embodiment, referring to FIGS. 13 and 3, a full-filment service according to the present invention may be performed as follows.

The communication interface unit 310 may receive order information for shipping goods from the online seller terminal 110 (S402).

The data extracting/classifying unit 320 may extract search data within the received order information and classify the extracted search data (S403).

The controller 340 may directly or control the data processing unit 330 to register a service keyword through artificial intelligence learning (S401), and compare the classified search data with the registered service keywords (S404). In addition, the control unit 340 generates full-filment service data related to a corresponding service keyword according to the comparison result (S405), and transmits the generated full-filment service data to the offline warehouse terminal 130 to It can be controlled to be released (S406).

According to another embodiment, referring to FIGS. 14 and 3, a full-filment service according to the present invention may be performed as follows.

The communication interface unit 310 may receive order information for shipping goods from the online seller terminal 110 (S502).

The data extracting/classifying unit 320 may extract search data within the received order information and classify the extracted search data (S503).

The control unit 340 may directly or control the data processing unit 330 to register a service keyword through artificial intelligence learning (S501), and compare the classified search data with the registered service keywords (S504/S505). . In addition, if there is no service keyword matching the comparison result, the control unit 340 generates a new service keyword (S506), and generates first full-filment service data related to the generated new service keyword (S507). , By transmitting the generated first full-filment service data to the offline warehouse terminal, it is possible to control the goods to be shipped (S508).

In the above, the service keyword registration process is continuously performed and updated, and may have been performed before the order information is received, and may be continuously updated after that.

In the above, the order information may include at least one of text data, image data, and audio data. A detailed description of the data processing process for each case will be described later.

Meanwhile, referring to FIG. 15, in the above, the control unit 340 transmits confirmation data including the generated first full-filment service data to the online seller terminal 110 (S601), and the online seller terminal When feedback data for the confirmation data is received from 110 (S602), the generated first full-filment service data is then transmitted to the offline warehouse terminal 130 to control the delivery of the corresponding article.

In the above, at least one of the first full-filment service data and the confirmation data is the new service keyword, information on the goods received in the offline warehouse 130 corresponding to the service keyword, and the quantity of goods orders extracted from the order information. It may include at least one or more of information, inventory information on the corresponding article, and total inventory information on articles received by the online seller 110 in the offline warehouse 130.

The controller 340 may control to register the generated new service keyword as a service keyword and update the service keyword database 350.

The controller 340 may parse the feedback data received from the online seller terminal 110 and identify the parsed feedback data content (S603).

If the content of the identified feedback data is a negative message (S604), the control unit 340 re-comparisons the search data re-extraction/reclassification and reclassified search data in the received order information with a registered service keyword ( S605), if there is no service keyword matching the recomparison result (S606), a new service keyword is generated (S607), and second full-filment service data related to the generated new service keyword is generated (S608), Alternatively, if there is a service keyword matching the re-comparison result, control may be performed to generate second full-filment service data according to the matching service keyword (S507).

As described above, the full-filment service server 120 accurately catches the products requested by the online seller 110 or its customers from the order information received from the online seller 110, and through the offline warehouse 120 The product is shipped or delivered to the customer.

In this case, the order information may be implemented in various forms, as described above, but may be a text data form, an image data form, another data form, or any combination of the aforementioned data forms. .

However, in the present specification, for convenience, in the case where the order information is made of text data and the case where the order information is made of image data, the contents of obtaining accurate logistics search data according to artificial intelligence learning from the two cases will be described.

First, when the order information is text data (including voice data), the processing or operation process of the full-filment service server 120 will be described as follows.

FIG. 4 may itself be a full-filment service server 120 according to an embodiment different from that of FIG. 3, or a detailed configuration of a configuration in the full-filment service server 120 shown in FIG. 3.

To aid in understanding of the present invention and for convenience of explanation, FIG. 4 is described below with a detailed configuration of the control unit 240 of the full-filment service server 120 of FIG. 3, but the scope of the present invention is not limited thereto. The control unit 240 automatically maps the goods of the offline warehouse 130 from the order information of the online seller 110 using an artificial intelligence-based search engine including contents to be described later.

Meanwhile, in FIG. 4, only core components are illustrated and described for convenience, but the present invention is not limited thereto.

Referring to FIG. 4, the full-filment service server 120 includes a communication module 310, a speech-to-text (STT) module 410, and a speech synthesis (Text-To-Speech; TTS). A module 420, a natural language processing module 430, an operation/comparison/control module 440, and the like.

The communication interface unit 310 may communicate with the database 350 or may communicate with the online seller 110 and the offline warehouse 130 described above. The communication interface unit 310 receives order information including text data from the online seller 110.

The communication interface unit 302 may receive information of the separate online seller 110 or/and the offline warehouse 130 in or together with the above-described order information. Here, the information of the online seller 110 is, for example, identification information (ID) for identifying the corresponding online seller 110, various status information (eg, physical state of the online seller terminal 110, online It may be a software and/or hardware state of the seller terminal 110, etc.). Meanwhile, the offline warehouse information may include information on the type or type of items in the warehouse, information on items from a specific seller, information on the quantity of each item, and other information on the state of the warehouse.

The STT module 410 may receive a voice input among order information received through the communication interface unit 310 and convert the received voice input into text data based on various known methods. The STT module 410 may generate a feature vector sequence by extracting features such as a keyword, such as a specific word, from a voice input.

The STT module 410 may generate a text recognition result, eg, a sequence of words, based on various models such as a deep neural network model. In this case, the STT module 410 may refer to the database 350 when processing the received voice input, that is, converting it into text data. but,

The natural language processing module 430 may receive a text input from the communication interface unit 310 or the STT module 410. Text input received from the natural language processing module 430 is, for example, from the text input received from the online seller 110 in the communication interface unit 310 or the voice input received from the communication interface unit 310 in the STT module 304 The generated text recognition result may be, for example, a sequence of words. The natural language processing module 430 may receive status information associated with a corresponding user input, for example, status information of the online seller 110 at the time of the user input, and the like, together with or after receiving the text input. As described above, the status information is, for example, various status information related to the online seller 110 at the time of user voice input and/or text input in the online seller 110 (eg, physical status of the online seller 110, software And/or hardware status, environment status information around the online seller 110, etc.).

The natural language processing module 430 may correspond the received text input to one or more service keywords. Here, the service keyword may be associated with a series of operation(s) that can be understood and performed by the operation/comparison/control module 440. The natural language processing module 430 may refer to the above-described state information in mapping the received text input to one or more service keywords. The natural language processing module 430 may refer to data of the database 350 to be described later in correlating the received text input to one or more service keywords.

The operation/comparison/control module 440 may include, for example, a predefined ontology model. The ontology model may be expressed as, for example, a hierarchical structure between nodes, and each node may be a service keyword or one of its sub-attributes. The ontology model may be dynamically changed by adding or deleting nodes in relation to a new service keyword, or by modifying a relationship between nodes.

The operation/comparison/control module 440 may be associated with words and/or phrases related to a service keyword or sub-attributes from text data, respectively. The operation/comparison/control module 440 may implement the ontology model in the form of, for example, a vocabulary dictionary (not shown) consisting of a set of nodes of a hierarchical structure and words and/or phrases associated with each node, The natural language processing module 430 may determine a service keyword based on the ontology model implemented in the form of a vocabulary dictionary as described above.

Upon receiving a text input or a sequence of words, the natural language processing module 430 may determine which nodes in the ontology model are associated with each word in the sequence, and determine a corresponding service keyword or attribute based on the determination. I can. Through this determination, the item corresponding to the service keyword can be identified.

When the natural language processing module 430 analyzes a text input based on an ontology model in the form of a vocabulary dictionary on the operation/comparison/control module 440, it may not be able to determine an appropriate service keyword. For example, the received text input does not include a predetermined word or phrase that is defined as being mapped to a specific service keyword, and by listing entity information that is completely irrelevant to a predetermined word or phrase, the desired service keyword is indirectly expressed. There may be cases, etc.

In other words, when the text input through the calculation/comparison/control module 440 is not direct product or quantity data, it is not easy to determine the service keyword through the natural language processing module 430.

To this end, the natural language processing module 430 to the calculation/comparison/control module 440 are learning basic data for machine learning, and include previous history, online dictionary data, direct input data, online seller data, and offline warehouse data. It can be used for the above decision by referring to data related to logistics.

Meanwhile, the service keyword determined in this way may be different from, for example, logistics-related data of an offline warehouse. These contents may be resolved by receiving related data in advance and referring to the dictionary definition of the service keyword, or processed in the form of a link.

For example, the text data is "Necklace Silver 1EA", "Necklace 1 Silver", "Necklace 1 Silver", "Early 1 Silver", "Neck Class 1 Silver", "Necklace 1ea Silver", It may be "one neck class silver", "one neck class silver", and so on. In this case, the natural language processing module 430 to the calculation/comparison/control module 440 finally, as the service keyword, accurately store the "one silver necklace" of the corresponding online seller 110 on the online warehouse 130. It can be determined by the article and its quantity.

The natural language processing module 430, on the other hand, determines an appropriate service keyword from the ontology model in the form of a vocabulary dictionary on the operation/comparison/control module 440, but when an error occurs, analyzes a plurality of entities present in the input. Based on this, an appropriate service keyword may be determined or a new service keyword may be generated and determined. For example, the natural language processing module 430 may determine that if an entity of the same type is repeated multiple times, such an input relates to a complaint or return request.

The natural language processing module 430 may secure various databases for words representing quality, such as size, color, performance, and price, among the received inputs, and refer to the service keyword determination.

The database 350 may store and manage history data including personalized data of the corresponding online seller 110. This may contribute to fast calculation and accurate processing/determination by referring to the processing or determination of the calculation/comparison/control module 440 to the natural language processing module 430 later. On the other hand, the personalization data included in the database 350 may include, for example, the seller's previous history, the seller's pronunciation, vocabulary preference, location, setting language, contact/friend list, and other personalization information of various sellers for each seller. I can.

As described above, the STT module 410 can obtain more accurate text data by referring to personalization data of each seller in the database 350, for example, pronunciation characteristics for each seller when converting the voice input into text data. have. In determining the service keyword, the natural language processing module 430 may determine a more accurate service keyword by referring to each seller personalization data of the database 350, for example, a personalization characteristic or context of each seller.

The database 350 may exist, for example, in the online seller terminal 110, or may be distributed and disposed in the online seller terminal 110 and the offline warehouse 130.

The operation/comparison/control module 440 may generate a series of operation flows corresponding to the service keyword determined by the natural language processing module 430. The operation/comparison/control module 440 may determine which operation and/or task to be performed in response to a service keyword received from the natural language processing module 430 and generate a detailed operation flow according thereto.

The TTS module 420 may receive a conversation response generated to be transmitted to the online seller terminal 110 by the conversation generating module 314. The conversation response received by the TTS module 420 may be a natural language having a text form or a sequence of words. The TTS module 420 may convert the received text input into a speech form according to various types of algorithms.

In addition, when determining each service keyword, the natural language processing module 430 to the operation/comparison/control module 440 may calculate a mapping or matching similarity and refer to it. Accordingly, when the calculated mapping or matching similarity is less than a predetermined threshold, the above-described process may be re-performed or returned to the online seller 110 to request confirmation, and may be determined accordingly. Meanwhile, the confirmation, that is, feedback of the online seller 110 may be reflected in the next calculation and keyword determination.

On the other hand, the natural language processing module 430 to the operation/comparison/control module 440 separately mark various text data, words, words, and associates that are not previously registered in the database 350, so that the mapping or matching similarity is applied. Can be reflected.

Next, an embodiment in which the full-filment service server 120 performs a full-filment service from an image is included in the order information received from the online seller 110 will be described.

5 to 7, a method for providing a full-filment service according to an embodiment of the present invention receives an image to be searched for a product from an online seller terminal 110.

The full-filment service server 120 classifies the received image of the online seller terminal 110 to be searched for products.

The full-filment service server 120 recognizes a product-related object from the classified online seller terminal 110's product search target image, and sets a product-related part in the object as a region of interest (ROI), The set region of interest (ROI) is cropped.

The full-filment service server 120 estimates the image similarity by calculating the cropped image data and the product image data stored in the database 350 using sub-networks included in a pre-learned network function. For example, the full-filment service server 120 performs a product search through estimation of an image-to-image method using a similarity estimation model for a portion of an ROI cropped from an image to be searched for a product.

The full-filment service server 120 transmits the corresponding product data to the terminal of the offline warehouse 130 as a result of the search based on the estimated image similarity.

The full-filment service server 120 communicates with the online seller terminal 110 shown in FIG. 1 to receive a variety of data from the online seller terminal 110, including product image data to be searched. Here, the image data to be searched received from the online seller terminal 110 may correspond to at least one of a captured or snapshot image, a captured image, a downloaded image, an image pre-stored in the online seller terminal 110, etc. have.

The full-filment service server 120 receives the received image data for product search of the online seller terminal 110.

The full-filment service server 120 classifies the image data to be searched for the received online seller terminal 110. Meanwhile, as described above, the full-filment service server 120 uses a deep learning-based multi-class classification model to classify the image to be searched for goods of the received online seller terminal 110 as described above. You can distinguish the types of. To this end, the full-filment service server 120 receives various types of product image data collected through the web, the product image of the manufacturer, the image provided by the online seller 110, and so on, through the deep learning model in advance. You can learn. The learned result may be stored in the database 350. Meanwhile, a deep learning-based multi-class classification model used in the full-filment service server 120 may be stored in the database 350 in advance.

The full-filment service server 120 performs object detection and cropping using a Faster Convolutional Neural Network Features (R-CNN) model. FIG. 5A shows a feature map pyramid extracted from an FPN, and FIG. 5B shows an example of an R-CNN structure using a feature map pyramid extracted from an FPN. A detailed description of the Faster R-CNN model and a description of (a) and (b) of FIGS. 5 to be described later refer to known contents, and detailed description thereof will be omitted. However, not only the currently published R-CNN model but also the future developed and known R-CNN model is used in the process of detecting an object containing an item and automatically cropping the item-related item from the detection object in connection with the present invention. Can be.

First, the full-filment service server 120 is identified as a product image as a result of the above-described classification, and when the class is confirmed, the object related to the product image of the identified class in the search target image previously received from the terminal 110 Recognize.

In other words, in particular, a Faster R-CNN model that has been trained in advance by specializing in an item item is constructed, and by using the constructed model, among at least one or more objects in the image received from the online seller terminal 110, particularly related to the item item. Objects can be detected. In the case of the conventional object detection model, it was common to detect all objects included in the received image, but the present invention determines and detects only the objects necessary from the image as meaningful objects through an object detection model that is specialized for item items and learned in advance. By doing so, it is possible to increase the overall efficiency of the system by reducing the object detection time and computation time.

On the other hand, although not shown, for example, depending on the user's selection or setting, for example, it is a target that has been learned in advance by specializing only on the specific item(s) handled, excluding items that are not handled by the online seller 110 among the item items. By using an object detection model, the efficiency can be further improved.

Alternatively, product items may be randomly classified and multiple object detection models corresponding to the classified items may be used. In this case, the object detection model is applied sequentially or in parallel according to a criterion such as high user interest, high sales, or high priority due to the high level of sales or the type or number of products possessed, to determine objects related to the item items. Can be detected.

In addition, in the case of object detection, since the cropping process described later remains, it is not necessary to perform a complicated process such as critically classifying objects with item items, but to detect objects according to predetermined criteria and perform the cropping process. I can.

Meanwhile, when a plurality of objects related to an article item are detected as a result of object detection in the full-filment service server 120, it may be processed according to a setting. Here, according to the setting, for example, all objects detected to select a specific object among a plurality of objects in the cropping process are transferred to the cropping process, or a plurality of detection objects are determined in advance for convenience in the cropping process. Among them, it is possible to determine only the object containing the item item related to the desired item item and transfer it to the cropping process. On the other hand, it is preferable to group objects including item items belonging to the same classification among the detected objects and transmit them through a cropping process.

The full-filment service server 120 determines the product image part of the actually identified class from the recognized object as a region of interest, that is, a Region of Interest (ROI), and crops the corresponding part, that is, an ROI part, from the object. . As the object detection and cropping model used in the present invention, through the R-CNN model, the ROI is automatically cropped and can be used immediately for a corresponding article search described later.

In this way, a learning model specialized for an item item is used through the full-filment service server 120 according to the present invention, and through this, recognition accuracy is increased, and a number of cropped images for each object are stored in the image search step in the online seller terminal 110. It is possible to improve the user's convenience by inserting a single coordinated image and searching for a corresponding product without manually inputting it individually.

The above description was a process of performing object detection and cropping a part of an actual item item to be searched from the detected object from the ROI. However, the present invention is not necessarily limited thereto. For example, in the case of object detection, it is used for searching by cropping an item item to be searched from an image. Although not shown, detection of an object including the item item according to learning or through a pre-input item item The process can be omitted. That is, since the item item to be used for search is extracted from the image in the end, the detection process of the object including the object may be removed, and the process of detecting the item item may be substituted for it. By doing this, the detected item item can be cropped immediately and the processing speed or the amount of association can be reduced, thereby increasing the system efficiency.

The full-filment service server 120 calculates the similarity by comparing the cropped product image with the retained product image stored in the database 350 and image-to-image. The full-filment service server 120 may set the ranking of product candidates returned as a search result based on the similarity calculated as described above.

According to the present invention, the full-filment service server 120 calculates the cropped image data and the product image data stored in the database using sub-networks included in a pre-learned network function to estimate and search the image similarity. And search for the product through it. Here, the sub-networks may include a convolutional neural network (CNN) structure among deep neural network structures. As shown in FIG. 7, the sub-networks may constitute at least a part of a Siamese network. However, the present invention is not necessarily limited thereto. Meanwhile, as described above, a detailed description of a convolutional neural network (CNN), a Siamese network, and the like among the deep neural network structures will be omitted herein with reference to known techniques.

According to the present invention, when the image-to-image search method using the Siamese network shown in FIG. 7 is used, an index of similarity can be calculated numerically. Therefore, by arbitrarily adjusting or setting the calculated similarity index, the product search result value may be changed, and through this adjustment, it may be more suited to the intention or needs of the online seller 110.

In relation to this, in the present invention, a distribution ontology may be constructed through the above-described similarity estimation model and used for product search.

In FIG. 6, a learning model is shown, and attribute values of detected objects can be extracted through such a learning model. As an example, Google's Inception-V3 learning model is shown in FIG. 6, and related information is referred to, and a detailed description of the learning model is omitted herein.

For example, images having the same attribute as the detected attribute may be read from the database 350 and used as a search candidate group. After that, the ranking of the candidate group can be determined in connection with the similarity estimation model based on the Siamese network and determined/returned as the search result.

In addition, it is possible to use the online seller 110 as a filter function to perform detailed search by excluding/including items returned as a search result according to the detected attribute.

That is, conventionally, there has been a hassle of having to directly input accurate product attribute information through the online seller 110 or the offline warehouse 130 terminal, but in the present invention, the distribution ontology is built in advance and the attribute is automatically extracted from the input image. By tagging, not only can the service procedure be shortened, but also the online seller 110 can increase convenience.

On the other hand, between the online seller-full-filment service server-offline warehouse, the above-described process, that is, product search request-product search result determination/return, and selection of the online seller 110 or the offline warehouse 130 according to the provision of the search result. The efficiency can be further increased by reflecting it in learning and updating it.

If the full-filment service server 120 has information on an online seller or an offline warehouse terminal, a personalized product search and provision service customized to the terminal through learning is also possible. In this way, when personalization is performed, the factor values of the above-described various models can be more simply configured, efficiency can be further improved through learning, and results for desired products can be quickly provided.

Meanwhile, in relation to the present invention, the database 350 stores various data described above, such as received images, product information, and distribution ontology described above. Meanwhile, the database 350 does not necessarily have to be built in the computing device 120 (or the online seller terminal 110). This database 350 also stores data in the form of a cloud, and, if necessary, returns a uniform resource locator (url) to access the url returned from the full-filment service server 120 or the online seller terminal 110 to be related. It can also be built to be able to acquire data.

The control unit 340 controls the entire process of the full-filment service server 120 as a whole, and may appropriately control each configuration as necessary.

Meanwhile, in the case of the image comparison method described above, in general, it is possible to increase the accuracy of calculation or determination by collecting and referring to information in a variety of 3D forms on an article. For example, referring to images acquired for various fields of view or angles such as the front, rear, top, bottom, side, etc. of the article, it is possible to contribute to fast calculation and accurate determination in the process of determining the article.

Although not shown, in the case where the order information received from the online seller 110 in the full-filment service server 120 does not contain only text or images alone, but is combined, the above-described contents may be appropriately combined to solve the problem. Therefore, it can be seen that such a case is also included in the embodiments of the present invention. For example, if the data extracted from the order information and classified or classified by cropping is text, then the natural language processing method is used according to the text processing method, and if it is an image, the image processing method is used to obtain the final data from the processed data according to each method Full-filment service can be provided based on one data.

In addition, in the above case, if the information on the same item or the final full-filment service search data obtained therefrom do not match, repeat the process or full-filment service data on the items based on each different data. Can be generated and delivered to the online seller 110 to receive confirmation. Related data, such as experience acquired in this process, may be used as learning data of artificial intelligence-based data according to the present invention and may be updated later.

A full-filment service process in the full-filment service system 100 shown in FIG. 9 will be described with reference to FIGS. 10 and 11.

Referring to FIG. 10, the online seller 110 stocks items that can be sold to the offline warehouse 130 in advance (S201).

When a customer places an order for a specific product with the online seller 110 (S202), the online seller 110 uploads the order information to the full-filment service server 120 (S203).

The full-filment service server 120 converts the order information into a database and transmits the order information to the offline warehouse 130 (S204).

The offline warehouse 130 picks up, packages, and releases the corresponding goods based on the transmitted order information (S205/S206).

The offline warehouse 130 uploads the invoice information to the full-filment service server 120 after the goods are released (S207).

The full-filment service server 120 provides the uploaded product invoice information to the online seller 110 (S208).

Referring to FIG. 11, a detailed processing process for the CS processing and return operation after FIG. 10 will be described.

When a customer requests a CS processing or return operation to the online seller 110 (S301), the online seller 110 transmits this to the full-filment service server 120 (S302).

The full-filment service server 120 determines the customer's CS processing/return job request information and transmits it to the offline warehouse 130 (S303).

After confirming this, the offline warehouse 130 transmits the feedback information to the online seller 110 through the full-filment service server 120 (S304, S305). In this case, the feedback information may include various data such as a restocking request according to a lack of quantity of goods, confirmation of whether another product is released, and data on a product delivery process according to a transfer request.

In addition, the offline warehouse 130 releases other items and delivers them to the customer (S306).

The offline warehouse 130 returns information on the reshipment to the online seller 110 through the full-filment service server 120 (S307, S308).

Thereafter, the online seller 110 receives the customer's feedback again. If the feedback is positive, the process ends. However, if the feedback is negative, the above-described procedure may be repeated or another procedure may be performed. The above other procedures may be the import and compensation of outgoing goods.

Meanwhile, in the above-described process, the full-filment service server 120 can continuously manage and control the full-filment service by converting and updating various types of data into a database in the form shown in FIG. 12.

On the other hand, the above-described method can be written in a program that can be executed on a computer, and can be implemented in a general-purpose digital computer that operates the program using a computer-readable medium. In addition, the structure of the data used in the above-described method can be recorded on a computer-readable medium through various means. Computer-readable media storing executable computer code for performing various methods of the present invention include magnetic storage media (e.g., 　 ROM, 　 floppy disk, 　 hard disk, etc.), 　 optical reading medium (e.g., CD-ROM, DVD Etc.).

Those of ordinary skill in the technical field related to the embodiments of the present invention will appreciate that it may be implemented in a modified form without departing from the essential characteristics of the above description. Therefore, the disclosed methods should be considered from a descriptive point of view rather than a limiting point of view. The scope of the present invention appears in the claims rather than the detailed description of the invention, and all differences within the scope equivalent thereto should be construed as being included in the scope of the present invention.

Claims (19)

In a method of providing a full-filment service in a service server between an online seller and an offline warehouse,
Registering a service keyword through artificial intelligence learning;
Receiving order information for releasing goods from the online seller terminal;
Extracting search data in the received order information and classifying the extracted search data;
Comparing the classified search data with registered service keywords;
Generating first full-filment service data related to a corresponding service keyword according to the comparison result; And
Transmitting the generated first full-filment service data to the offline warehouse terminal to control the delivery of the corresponding article; and a full-filment service providing method in a service server comprising a. The method of claim 1,
The received order information,
A method for providing a full-filment service in a service server, characterized in that at least one of text data, image data, and audio data is included. The method of claim 1,
Transmitting the generated first full-filment service data to the offline warehouse terminal to control the goods to be shipped,
Full-filment service in a service server, characterized in that after the confirmation data including the generated first full-filment service data is transmitted to the online seller terminal and feedback data for the confirmation data is received from the online seller terminal. Delivery method. The method of claim 3,
At least one of the first full-filment service data and the confirmation data,
The new service keyword, information on the goods received in the offline warehouse corresponding to the service keyword, information on the quantity of goods orders extracted from the order information, inventory information on the corresponding goods, and goods stocked by the online seller to the offline warehouse Full-filment service providing method in a service server comprising at least one or more of the total inventory information for. In a method of providing a full-filment service in a service server between an online seller and an offline warehouse,
Registering a service keyword through artificial intelligence learning;
Receiving order information for releasing goods from the online seller terminal;
Extracting search data in the received order information, classifying the extracted search data, and comparing the classified search data with registered service keywords;
If there is no service keyword matching the comparison result, generating a new service keyword and generating first full-filment service data related to the generated new service keyword; And
Transmitting the generated first full-filment service data to the offline warehouse terminal to control the delivery of the corresponding article; and a full-filment service providing method in a service server comprising a. The method of claim 5,
Transmitting the generated first full-filment service data to the offline warehouse terminal to control the goods to be shipped,
Full-filment in a service server, characterized in that after the confirmation data including the generated first full-filment service data is transmitted to the online seller terminal, and feedback data for the confirmation data is received from the online seller terminal. How to provide the service. The method according to claim 5 or 6,
And registering the generated new service keyword as a service keyword and updating a service keyword database. The method of claim 7,
The step of registering the new service keyword,
When the confirmation data including the generated first full-filment service data is transmitted to the online seller terminal, and when feedback data for the confirmation data is received from the online seller terminal, it is performed based on the received feedback data. How to provide full-filment service in the service server The method of claim 8,
Parsing the feedback data received from the online seller terminal; And
And identifying the parsed feedback data content. The method of claim 9,
If the identified feedback data content is a negative message,
Extracting/classifying search data in the received order information and re-comparing the classified search data with registered service keywords,
If there is no service keyword matching the recomparison result, generating a new service keyword and generating second full-filment service data related to the generated new service keyword, or
And generating second full-filment service data according to the matched service keyword if there is a matching service keyword in the re-comparison result. In a service server that provides a full-filment service between an online seller and an offline warehouse,
A communication interface unit for receiving order information for discharging goods from the online seller terminal;
A data extraction/classification unit for extracting search data in the received order information and classifying the extracted search data; And
Registering a service keyword through artificial intelligence learning, comparing the classified search data with the registered service keyword, generating first full-filment service data related to the service keyword according to the comparison result, and generating the first full-filment service data A service server comprising a; a control unit that transmits full-filment service data to the offline warehouse terminal to control the delivery of the corresponding article. The method of claim 11,
The received order information,
A service server comprising at least one of text data, image data, and audio data. The method of claim 11,
The control unit,
After the confirmation data including the generated first full-filment service data is transmitted to the online seller terminal and feedback data for the confirmation data is received from the online seller terminal, the generated first full-filment service data is Service server, characterized in that for controlling the goods to be shipped by transmitting them to the offline warehouse terminal. The method of claim 13,
At least one of the first full-filment service data and the confirmation data,
The new service keyword, information on the goods received in the offline warehouse corresponding to the service keyword, information on the quantity of goods orders extracted from the order information, inventory information on the corresponding goods, and goods stocked by the online seller to the offline warehouse A service server comprising at least one or more of total inventory information for. In a service server that provides a full-filment service between an online seller and an offline warehouse,
A communication interface unit for receiving order information for discharging goods from the online seller terminal;
A data extraction/classification unit for extracting search data in the received order information and classifying the extracted search data; And
A service keyword is registered through artificial intelligence learning, the classified search data is compared with the registered service keyword, and if there is no service keyword matching the comparison result, a new service keyword is generated, and the generated new service keyword and A service server comprising: a control unit for generating related first full-filment service data, transmitting the generated first full-filment service data to the offline warehouse terminal, and controlling the goods to be shipped. The method of claim 15,
The control unit,
After the confirmation data including the generated first full-filment service data is transmitted to the online seller terminal and feedback data for the confirmation data is received from the online seller terminal, the generated first full-filment service data The service server, characterized in that for controlling so that the goods are shipped by transmitting them to the offline warehouse terminal. The method of claim 15 or 16,
The control unit,
And registering the generated new service keyword as a service keyword and controlling to update a service keyword database. The method of claim 17,
The control unit,
When the confirmation data including the generated first full-filment service data is transmitted to the online seller terminal, and when feedback data for the confirmation data is received from the online seller terminal, based on the received feedback data, the new Service server, characterized in that for registering a service keyword. The method of claim 18,
The control unit,
Parsing the feedback data received from the online seller terminal, and identifying the parsed feedback data content.

Images