Interactive order processing method and system including artificial neural network model

Abstract

An interactive order processing system including an artificial neural network model, according to an embodiment of the present disclosure, comprises: a product information management module including a product vector generation unit that generates a product vector on the basis of product information and a product database that stores the product vector; and an order processing module that performs a product search in the product database on the basis of order data. The product vector generation unit stores, in the product database, structured data extracted from the product information.

Description

Interactive order processing method and system including an artificial neural network model

[0001] The present disclosure relates to a method and system for interactive order processing, and more particularly, to a method and system for interactive order processing using a vector database and a large-scale language model generated based on product information.

[0002] Kiosks are computerized automated systems installed in public spaces for customer convenience. They are used for a variety of purposes, from menu ordering at restaurants and cafes to ticketing at sports stadiums, amusement parks, movie theaters, terminals, and stations. With the recent rise of unmanned automated systems, their use has rapidly increased. However, the complex menu selection process at kiosks can actually increase the time required to order, making them difficult for children and the elderly to use. To utilize kiosks and unmanned automated systems, a system that is more convenient for users and provides an efficient ordering process is needed.

[0003] The present disclosure is conceived in response to the aforementioned background technology, and aims to provide an interactive unmanned ordering system.

[0004] This disclosure can efficiently transform the cumbersome ordering process and provide a new experience to customers by utilizing a large language model (LLM).

[0005] However, the problems to be solved in this disclosure are not limited to the problems mentioned above, and other problems not mentioned can be clearly understood based on the description below.

[0006] An interactive order processing system including an artificial neural network model according to one embodiment of the present disclosure for realizing the aforementioned task includes a product vector generation unit for generating a product vector based on product information, a product information management module including a product database storing the product vector, and an order processing module for performing a product search in the product database based on order data. The product vector generation unit is characterized in that it stores structured data extracted from the product information in the product database.

[0007] Alternatively, the product database is characterized by including a structured database and a vector database.

[0008] Alternatively, the structured data may include a price, a product name, or a spiciness of the food.

[0009] Alternatively, the order processing module is characterized by including a first embedding model that generates a query vector based on the order data, and a first sub-unit that performs a similarity judgment in the vector database based on the query vector and receives unstructured search data.

[0010] Alternatively, the order processing module is characterized in that it performs a structured search in the structured database based on the order data and receives structured search data.

[0011] Alternatively, the product information management module includes product service information,

[0012] The above product service information is characterized by including at least one of industry, provided services, operating rules, or operating hours.

[0013] Alternatively, the order processing module is characterized in that it generates order processing data based on the structured search data, the unstructured search data, and the product service information.

[0014] Alternatively, the order processing module includes an answer generation module that generates an answer to the order data through a first neural network model based on the structured search data, wherein the first neural network model is one of Large Language Models (LLM), and the answer generation module is characterized in that it generates an answer through Retrieval-Augmented Generation (RAG).

[0015] According to one embodiment of the present disclosure, an interactive order processing method performed by a computing device including at least one processor includes the steps of performing preprocessing on input data to generate order data, performing a structured search on a product database based on the order data and generating structured search data, generating a query vector based on the order data, and performing a similarity judgment on a product database using the query vector as input and generating unstructured search data.

[0016] Alternatively, the product database is characterized by including a vector database that stores product vectors generated through word embedding based on preset product information.

[0017] Alternatively, the step of generating the structured search data comprises the step of generating a structured search query for one or more of price, product name, or spiciness of food.

[0018] Alternatively, the step of generating the unstructured search data comprises one of pronunciation recognition, language translation, or matching menu recommendation.

[0019] Alternatively, the method comprises a step of generating an answer to the order data through a first neural network model based on the structured search data, wherein the first neural network model is one of large-scale language models, and the answer is generated through Retrieval-Augmented Generation (RAG).

[0020] According to one embodiment of the present disclosure, a computing device for performing interactive order processing includes a processor including at least one core, a memory including program codes executable by the processor, and a network unit for acquiring data. The processor performs preprocessing on input data to generate order data, performs a structured search on a product database based on the order data to generate structured search data, generates a query vector based on the order data, performs a similarity judgment on the product database using the query vector as input, and generates unstructured search data. The product database is characterized in that it includes a vector database that stores a product vector generated through word embedding based on preset product information.

[0021] The present disclosure can predict products desired by customers and effectively process orders by using a vector database generated based on product information.

[0022] The present disclosure utilizes an artificial neural network model and a vector database to reflect additional conditions to orders entered in a conversational manner and to more accurately predict orders.

[0023] This disclosure enables processing of orders in a natural conversational format and application of personalized processes to each user, taking into account language, constitution, physical conditions, etc.

[0024] FIG. 1 is a block diagram of a computing device according to one embodiment of the present disclosure.

[0025] FIG. 2 is a block diagram illustrating an interactive order processing system according to one embodiment of the present disclosure.

[0026] FIG. 3 is a block diagram illustrating an interactive order processing system according to one embodiment of the present disclosure.

[0027] FIG. 4 is a diagram illustrating the operation of a product information management module and an order processing module according to one embodiment of the present disclosure.

[0028] FIG. 5 is a diagram illustrating word embedding according to one embodiment of the present disclosure.

[0029] FIG. 6 is a flowchart illustrating an interactive order processing method according to one embodiment of the present disclosure.

[0030] Below, embodiments of the present disclosure are described in detail with reference to the attached drawings so that those skilled in the art can easily implement the present disclosure. The embodiments presented in this disclosure are provided to enable those skilled in the art to utilize or implement the contents of the present disclosure. Accordingly, various modifications to the embodiments of the present disclosure will be apparent to those skilled in the art. That is, the present disclosure may be implemented in various different forms and is not limited to the embodiments described below.

[0031] Throughout the specification of this disclosure, identical or similar drawing numbers refer to identical or similar components. Furthermore, for the purpose of clearly describing the disclosure, drawing numbers for parts in the drawings that are not relevant to the description of the disclosure may be omitted.

[0032] The term "or" as used herein is intended to mean an inclusive "or" rather than an exclusive "or." That is, unless otherwise specified herein or clear from context, "X employs A or B" should be understood to mean either of the natural inclusive permutations. For example, unless otherwise specified herein or clear from context, "X employs A or B" can be interpreted to mean either X employs A, X employs B, or X employs both A and B.

[0033] The term "and / or" as used herein should be understood to refer to and include all possible combinations of one or more of the related concepts listed.

[0034] The terms "comprises" and / or "comprising" as used herein should be understood to mean the presence of certain features and / or components. However, it should be understood that the terms "comprises" and / or "comprising" do not exclude the presence or addition of one or more other features, other components, and / or combinations thereof.

[0035] Unless otherwise specified in this disclosure or unless the context makes it clear that the singular form is intended to be referred to, the singular should generally be construed to include “one or more.”

[0036] The term "Nth (N is a natural number)" used in the present disclosure can be understood as an expression used to mutually distinguish components of the present disclosure based on a predetermined standard such as a functional perspective, a structural perspective, or convenience of explanation. For example, components performing different functional roles in the present disclosure can be distinguished as a first component or a second component. However, components that are substantially the same within the technical spirit of the present disclosure but must be distinguished for convenience of explanation may also be distinguished as a first component or a second component.

[0037] The term "acquisition" as used in this disclosure may be understood to mean not only receiving data through a wired or wireless communication network with an external device or system, but also generating data in an on-device form.

[0038] Meanwhile, the term "module" or "unit" used in the present disclosure can be understood as a term referring to an independent functional unit that processes computing resources, such as a computer-related entity, firmware, software or a part thereof, hardware or a part thereof, or a combination of software and hardware. At this time, the "module" or "unit" may be a unit composed of a single element, or a unit expressed as a combination or set of multiple elements. For example, as a narrow concept, a "module" or "unit" may refer to a hardware element of a computing device or a set thereof, an application program that performs a specific function of software, a processing process implemented through software execution, or a set of instructions for program execution, etc. In addition, as a broad concept, a "module" or "unit" may refer to the computing device itself that constitutes the system, or an application running on the computing device, etc. However, since the above-described concept is only an example, the concept of “module” or “part” may be defined in various ways within a range understandable to those skilled in the art based on the contents of the present disclosure.

[0039] The term "model" as used herein may be understood as a system implemented using mathematical concepts and language to solve a specific problem, a set of software units to solve a specific problem, or an abstract model of a processing process to solve a specific problem. For example, a neural network "model" may refer to the entire system implemented as a neural network that has problem-solving capabilities through learning. In this case, the neural network can have problem-solving capabilities by optimizing the parameters connecting nodes or neurons through learning. A neural network "model" may include a single neural network or a set of neural networks that are a combination of multiple neural networks.

[0040] The term "block" used in this disclosure can be understood as a set of configurations categorized based on various criteria, such as type and function. Therefore, the configurations classified as a single "block" can vary depending on the criteria. For example, a neural network "block" can be understood as a set of neural networks including at least one neural network. In this case, the neural networks included in the neural network "block" can be assumed to perform specific operations identically.

[0041] The term "operation function" used in this disclosure can be understood as a mathematical expression for a component that performs a specific function or processes an operation. For example, the "operation function" of a neural network block can be understood as a mathematical expression representing a neural network block that processes a specific operation. Accordingly, the relationship between the input and output of a neural network block can be expressed as a formula through the "operation function" of the neural network block.

[0042] The explanation of the above terms is intended to aid understanding of the present disclosure. Therefore, unless explicitly stated as limiting the contents of the present disclosure, it should be noted that the above terms are not intended to limit the technical ideas of the present disclosure.

[0043] FIG. 1 is a block diagram of a computing device according to one embodiment of the present disclosure.

[0044] A computing device (100) according to one embodiment of the present disclosure may be a hardware device or a part of a hardware device that performs comprehensive processing and calculation of data, or may be a software-based computing environment connected to a communication network. For example, the computing device (100) may be a server that performs intensive data processing functions and shares resources, or may be a client that shares resources through interaction with a server. In addition, the computing device (100) may be a cloud system in which multiple servers and clients interact to comprehensively process data. Since the above description is only one example related to the type of computing device (100), the type of computing device (100) may be configured in various ways within a range understandable to those skilled in the art based on the contents of the present disclosure.

[0045] Referring to FIG. 1, a computing device (100) according to one embodiment of the present disclosure may include a processor (110), a memory (120), and a network unit (130). However, FIG. 1 is merely an example, and the computing device (100) may include other components for implementing a computing environment. In addition, only some of the disclosed components may be included in the computing device (100).

[0046] The processor (110) according to one embodiment of the present disclosure may be understood as a configuration unit including hardware and / or software for performing computing operations. For example, the processor (110) may read a computer program to perform data processing for machine learning. The processor (110) may process computational processes such as processing of input data for machine learning, feature extraction for machine learning, embedding, error calculation based on backpropagation, etc. The processor (110) for performing such data processing may include a central processing unit (CPU), a general purpose graphics processing unit (GPGPU), a tensor processing unit (TPU), an application specific integrated circuit (ASIC), or a field programmable gate array (FPGA). The type of the processor (110) described above is only one example, and thus, the type of the processor (110) may be configured in various ways within a range understandable to those skilled in the art based on the contents of the present disclosure.

[0047] The memory (120) according to one embodiment of the present disclosure may be understood as a configuration unit including hardware and / or software for storing and managing data processed in the computing device (100). That is, the memory (120) may store any type of data generated or determined by the processor (110) and any type of data received by the network unit (130). For example, the memory (120) may include at least one type of storage medium among a flash memory type, a hard disk type, a multimedia card micro type, a card type memory, a random access memory (RAM), a static random access memory (SRAM), a read-only memory (ROM), an electrically erasable programmable read-only memory (EEPROM), a programmable read-only memory (PROM), a magnetic memory, a magnetic disk, and an optical disk. In addition, the memory (120) may also include a database system that controls and manages data in a predetermined system. The type of memory (120) described above is only one example, and thus the type of memory (120) can be configured in various ways within a range understandable to those skilled in the art based on the contents of the present disclosure.

[0048] The memory (120) can structure and organize and manage data, combinations of data, and program codes executable by the processor (110) required for the processor (110) to perform operations. For example, the memory (120) can store data received through the network unit (130) described below. The memory (120) can store program codes that operate a neural network model to receive data as input and perform learning, program codes that operate a neural network model to receive data as input and perform inference according to the purpose of use of the computing device (100), and processed data generated as the program codes are executed.

[0049] The network unit (130) according to one embodiment of the present disclosure may be understood as a component that transmits and receives data through any type of known wired or wireless communication system. For example, the network unit (130) may perform data transmission and reception using a wired or wireless communication system such as a local area network (LAN), wideband code division multiple access (WCDMA), long term evolution (LTE), wireless broadband internet (WiBro), fifth generation mobile communication (5G), ultra wide-band, ZigBee, radio frequency (RF) communication, wireless LAN, wireless fidelity, near field communication (NFC), or Bluetooth. Since the above-described communication systems are only examples, the wired and wireless communication system for data transmission and reception of the network unit (130) may be applied in various ways other than the above-described examples.

[0050] The network unit (130) can receive data necessary for the processor (110) to perform calculations through wired or wireless communication with any system or any client, etc. In addition, the network unit (130) can transmit data generated through calculations of the processor (110) through wired or wireless communication with any system or any client, etc. For example, the network unit (130) can receive data through communication with an embedded model or a large-scale language model, a database, a cloud server, a computing device, etc. The network unit (130) can transmit output data of a neural network model, intermediate data derived from the calculation process of the processor (110), processed data, etc. through communication with the aforementioned database, server, or computing device.

[0051]

[0052] FIG. 2 is a block diagram illustrating an interactive order processing system according to one embodiment of the present disclosure.

[0053] Referring to FIG. 2, the interactive order processing system (200) may include an order processing module (210), a product information management module (220), an input module (230), and an output module (240).

[0054] The order processing module (210) can request a search for similar products from the product information management module (220) based on order data generated from the input module (230) or the preprocessing unit. The order processing module (210) can combine the order data (orders received from the input module (230) or the preprocessing unit) and the search results of the product information management module (220) to request processing from the interactive artificial intelligence system (not shown) and receive a response. The order processing module (210) can transmit the response generated by the interactive artificial intelligence system to the output module (240) or the output conversion unit, or to a separate user terminal for order processing. The order processing module (210) can record the order processing result. The interactive artificial intelligence system can generate a response based on the order and context (situation, context) transmitted from the order processing module (210). The interactive artificial intelligence system can include natural language processing (NLP) and a large-scale language model.

[0055] The product information management module (220) can convert product or service information provided by a user into a vector capable of similarity search. The product information management module (220) can store the converted vector in a vector database that provides a similarity search function. The product information management module (220) can process a similarity search request transmitted from the order processing module (210).

[0056] The input module (230) may include one or more of a handwriting recognition device, a touchscreen, a microphone, or a voice recognition device for receiving an order from a user. The input module (230) may further include a preprocessing unit for converting input data into characters, such as an STT (Sound To Text) processing unit or an OCR (Optical Character Recognition) processing unit.

[0057] The output module (240) may include one or more of a display and a speaker. The output module (240) may further include an output conversion unit for transmitting the response generated by the order processing module (210) to the user. For example, the output conversion unit may be a TTS (Text To Sound) processing unit for responding to the user through voice, or a character processing unit for displaying order details on the display.

[0058] In Fig. 2, the interactive order processing system (200) is described by dividing it into an order processing module (210), a product information management module (220), an input module (230), and an output module (240), but this is for convenience of explanation and is not limited thereto, and the method of dividing the modules or the inclusion relationship can be implemented in various forms.

[0059]

[0060] FIG. 3 is a block diagram illustrating an interactive order processing system according to one embodiment of the present disclosure.

[0061] Referring to FIG. 3, the interactive order processing system may include an order processing module (310), a product information management module (320), a first user terminal (380), and a second user terminal (370).

[0062] The first user terminal (380) may be a kiosk or ordering terminal used for ordering. The first user terminal (380) may collect order information or order-related questions from customers or users ordering products. The first user terminal (380) may include various components for a user interface (UI), such as a handwriting input / output unit (382) and a voice input / output unit (384). The interactive order processing system may include an input conversion module (330) for processing input data collected from the first user terminal (380) and an output conversion module (350) for converting output data.

[0063] The input conversion module (330) may further include a preprocessing unit for converting input data into characters, such as an STT (Sound To Text) processing unit (334) and an OCR (Optical Character Recognition) processing unit (332).

[0064] The output conversion module (350) may include a TTS (Text To Sound) processing unit for responding to the user through voice, and a character processing unit for displaying order details on the display.

[0065] The order processing module (310) may include a product search unit (312), a first embedding model (313), an answer generation unit (314), a first neural network model (315), and an order transmission unit (316).

[0066] The product search unit (312) can perform a product information search when order data is received. The product search unit (312) can generate a structured search query based on the order data and collect structured data from the product vector database (324) using the generated structured search query. The first embedding model (313) can generate a query vector based on the order data. The product search unit (312) can perform a similarity determination in the vector database (424) based on the generated query vector. The product search unit (312) can transmit the similarity determination result or unstructured search data to the answer generation unit (313).

[0067] The response generation unit (314) can generate responses using structured data and unstructured search data received from the received order data and product information management module (320). The response generation unit (314) can generate responses using the first neural network model (315). The first neural network model (315) can be a transformer-based model such as GPT, BERT, BART, or T5, or a large-scale language model (LLM).

[0068] The order transmission unit (316) can transmit the final order or response generated by the order processing module (310) to the service provider or the second user terminal (370). The order transmission unit (316) can receive and record order processing process or order processing result information from the service provider or the second user terminal (370).

[0069] The product information management module (320) may include a product vector generation unit (322) and a product vector database (324). The product vector generation unit (322) may generate a product vector based on product information. The product vector generation unit (322) may extract structured or unstructured product information data in the form of documents or text data. The product vector generation unit (322) may utilize a trained large-scale language model or neural network model to extract structured or unstructured data. The product vector generation unit (322) may store the extracted structured data in the product vector database (324). The product information management module (320) may store the extracted structured data as metadata in the product vector database (324) or in a separate structured database. The product vector generation unit (322) may embed the extracted unstructured data to generate a vector. The product vector generation unit (322) may store the embedded vector in the product vector database (324).

[0070] The second user terminal (370) may be a terminal used by the service provider to confirm an order and provide menus, products, and services corresponding to the order. The second user terminal (370) may transmit information collected from the first user terminal (380) or results processed by the order processing module (310) to the service provider.

[0071]

[0072] FIG. 4 is a diagram illustrating the operation of a product information management module and an order processing module according to one embodiment of the present disclosure.

[0073] Referring to FIG. 4, the interactive order processing system may include a product information management module (400) and an order processing module (500). The product information management module (400) may include a product vector generation unit (410) and a product database (420).

[0074] The product vector generation unit (410) can generate a product vector based on product information. The product vector generation unit (410) can include a fourth sub-unit (412) and a first embedding model (414). The fourth sub-unit (412) can receive product information in the form of a document or text data (S401). The fourth sub-unit (412) can extract structured data or unstructured data from the input product information (S402). The fourth sub-unit (412) can utilize a large-scale language model or neural network model trained to extract structured data or unstructured data. The first embedding model (414) can be a neural network model for converting the input product information into a vector. The product vector generation unit (410) can store the structured data extracted by the fourth sub-unit (412) in a structured database (422) (S403). The first embedding model (414) can create a vector by embedding the unstructured data extracted from the fourth sub-unit (412) (S404).

[0075] The product database (420) may include a structured database (422) and a vector database (424). The interactive order processing system may store structured data extracted from product information in the structured database (422). The interactive order processing system may convert unstructured data extracted from product information into vectors and store them in the vector database (424). The interactive order processing system may also store structured data as metadata in the vector database (424) rather than in a separate structured database (422). The product information management module (400) may store and utilize various information related to rules and industries related to orders in addition to structured and unstructured data extracted from product information.

[0076] The product information management module (400) may further include a first embedding model (530), a first sub-unit (540), a second sub-unit (510), and a third sub-unit (520) used for searching a product database (420).

[0077] The order processing module (500) can perform a search for related product information when order data is received (S501). The second sub-unit (510) can generate a structured search query based on the order data (S502). The third sub-unit (520) can search for structured data in the structured database (422) using the generated structured search query and transmit the retrieved structured search data to the order processing module (500) (S504). The first embedding model (530) can generate a query vector based on the order data (S503). The first sub-unit (540) can perform a similarity determination in the vector database (424) based on the generated query vector. The first sub-unit (540) can transmit the similarity determination result or unstructured search data to the order processing module (500) (S503).

[0078] The order processing module (500) can accurately and quickly process and respond to orders by considering structured data such as price, product name, or spiciness of food searched from the structured database (422) and the results of similarity judgment performed from the vector database or unstructured search data, and can provide more accurate answers by considering the industry, provided services, operating rules, operating hours, or general order criteria stored in the product information management module (400). In Fig. 4, the functions required for the interactive order processing process are described by dividing them into modules and units, but they are not limited thereto and can be implemented in various forms.

[0079]

[0080] FIG. 5 is a diagram illustrating word embedding according to one embodiment of the present disclosure.

[0081] Referring to Figure 5, the interactive order processing system can generate vectors using word embeddings based on product information. During the training process of a deep learning model, the interactive order processing system can convert unstructured data, such as language and images, into meaningful numbers (or vectors).

[0082] An interactive order processing system or product information management module can represent words contained in product information as n-dimensional vectors with n features. The interactive order processing system can display these vector-represented words in a two-dimensional space through dimensionality reduction and determine similarity based on the distance between vectors.

[0083] For example, a conversational order processing system can recognize similar pronunciations, such as correcting the pronunciation of 'tangsuyok' to 'tangsuyuk' by judging the similarity between the first word (W1) included in product information and the second word (W2) included in order data. A conversational order processing system can recognize different languages, such as recognizing 'Soju' as 'soju' by judging the similarity between the first word (W1) included in product information and the second word (W2) included in order data. A conversational order processing system can recommend menus that match the weather, such as matching 'rainy day' with 'makgeolli' by judging the similarity between the first word (W1) included in product information and the second word (W2) included in order data. A conversational order processing system can recommend menus that match, such as 'makgeolli' and 'pajeon', by judging the similarity between the first word (W1) included in product information and the second word (W2) included in order data.

[0084]

[0085] FIG. 6 is a flowchart illustrating an interactive order processing method according to one embodiment of the present disclosure.

[0086] Referring to Figure 6, the interactive order processing system can perform input data preprocessing (S110). The interactive order processing system can convert input data collected through a kiosk or first user terminal, such as a customer's voice or handwriting. The interactive order processing system can convert the voice or handwriting into text to generate order data. The interactive order processing system can separate third-party conversations, conversations unrelated to the order, and various noises from the order data.

[0087] The interactive order processing system performs a product search step (S120). The interactive order processing system can search for products in a product database based on order data. The product database can include a structured database and a vector database. The product database can store structured information in the vector database in the form of metadata. The interactive order processing system can perform structured searches for preset items, such as price and product name, in the structured database. Structured searches enable the interactive order processing system to quickly process information with fixed values, such as price and product name. The interactive order processing system can quickly compare similarities between vectors in the vector database, which stores embedded vectors. The interactive order processing system can use the vector database to quickly search for sentences or paragraphs that can answer a customer's question. The interactive order processing system can utilize a trained large-scale language model or neural network model to extract structured and unstructured data.

[0088] For example, during the product ordering process, a customer can provide both structured information, such as price, and unstructured information, such as smell and texture. For example, a customer might submit a request to a conversational order processing system, such as "Recommend a cool dish under 10,000 won," which includes the unstructured information "cool" and the structured information "10,000 won." The conversational order processing system can process structured information, such as price, from a structured database or metadata, and unstructured information, such as smell and texture, from a vector database.

[0089] The interactive order processing system performs a response generation step (S130).

[0090] A conversational order processing system can quickly extract potential answers from a product database containing product information, such as menus, and then generate answers to questions using a large-scale language model based on the extracted information. A conversational order processing system can generate answers using the Retrieval Augmented Generation (RAG) method.

[0091] A conversational order processing system can retrieve the most relevant phrases or phrases that might contain answers to an order during the order processing process. A conversational order processing system can retrieve documents contextually similar to the question from a database that stores data related to the answers, using a model such as Dense Passage Retrieval (DPR).

[0092] Additionally, conversational order processing systems can generate context-appropriate responses from retrieved documents. Using context-sensitive language provided by large-scale language models, conversational order processing systems can synthesize received information and generate responses appropriate to questions or requests arising during the ordering process. For example, conversational order processing systems can utilize Transformer-based models such as GPT, BERT, BART, or T5.

[0093] The conversational order processing system can convert the generated responses into voice and output them via an output module, a first user terminal, or a kiosk. In addition to voice, the conversational order processing system can also output the responses via a user interface included in the first user terminal or kiosk.

[0094] The interactive order processing system performs the order delivery step (S140). Once an order is confirmed based on the order data received from the order processing module, the interactive order processing system can deliver the order to the service provider or a second user terminal. The interactive order processing system can record the availability of menus, products, and services corresponding to the order and provide guidance to the customer via the first user terminal.

[0095] The various embodiments of the present disclosure described above can be combined with additional embodiments and modified within the scope understood by those skilled in the art in light of the detailed description above. It should be understood that the embodiments of the present disclosure are illustrative in all respects and not restrictive. For example, each component described as a single component may be implemented in a distributed manner, and likewise, components described as distributed may be implemented in a combined manner. Accordingly, all changes or modifications derived from the meaning, scope, and equivalent concepts of the claims of the present disclosure should be construed as being included within the scope of the present disclosure.

[0096]

Claims

1. In an interactive order processing system including an artificial neural network model, A product information management module including a product vector generation unit that generates a product vector based on product information and a product database that stores the product vector; and An order processing module that performs a product search in the product database based on order data; The above product vector generation unit is characterized in that it stores the structured data extracted from the product information in the product database. Interactive order processing system.

2. In paragraph 1, The above product database is characterized in that it includes a structured database and a vector database. Interactive order processing system.

3. In paragraph 1, The above structured data is characterized in that it includes price, product name or spiciness of food. Interactive order processing system.

4. In paragraph 2, The order processing module is characterized in that it includes a first embedding model that generates a query vector based on the order data, and a first sub-unit that performs a similarity judgment in the vector database based on the query vector and receives unstructured search data. Interactive order processing system.

5. In paragraph 4, The above order processing module is characterized in that it performs a structured search in the structured database based on the order data and receives structured search data. Interactive order processing system.

6. In paragraph 5, The above product information management module includes product service information, The above product service information is characterized in that it includes at least one of the following: industry, provided services, operating rules, or operating hours. Interactive order processing system.

7. In paragraph 6, The above order processing module is characterized in that it generates order processing data based on the structured search data, unstructured search data, and product service information. Interactive order processing system.

8. In paragraph 7, The above order processing module includes an answer generation module that generates an answer to the order data through a first neural network model based on the structured search data, The first neural network model is one of the Large Language Models (LLM), and the answer generation module is characterized in that it generates an answer through Retrieval-Augmented Generation (RAG). Interactive order processing system.

9. An interactive order processing method performed by a computing device including at least one processor, A step of generating order data by performing preprocessing on input data; A step of performing a structured search in a product database based on the above order data and generating structured search data; A step of generating a query vector based on the above order data; and A method characterized by comprising: performing a similarity judgment on a product database using the above query vector as input and generating unstructured search data; method.

10. In paragraph 9, The above product database is characterized in that it includes a vector database that stores a product vector generated through word embedding based on preset product information. method.

11. In paragraph 10, The step of generating the above structured search data includes the step of generating a structured search query for at least one of price, product name, or spiciness of food. method.

12. In paragraph 11, The step of generating the above unstructured search data is characterized in that it includes any one of similar pronunciation recognition, language translation, or matching menu recommendation. method.

13. In paragraph 10, A step of generating an answer to the order data through a first neural network model based on the above-mentioned structured search data is included, The first neural network model is one of the large-scale language models, and the answer is characterized in that it is generated through Retrieval-Augmented Generation (RAG). method.

14. A computing device that performs interactive order processing, A processor comprising at least one core; A memory containing program codes executable by the processor; and Network unit for acquiring data; Including, The above processor, Performing preprocessing on input data to generate order data, performing a structured search on a product database based on the order data to generate structured search data, generating a query vector based on the order data, performing a similarity judgment on a product database using the query vector as input, and generating unstructured search data. The above product database is characterized in that it includes a vector database that stores a product vector generated through word embedding based on preset product information. Computing device.

Images