On-device based system and method for preventing leakage of personal information and providing personalized support

Abstract

An on-device-based personal information leakage prevention and personalized response system according to one embodiment of the present invention includes a plurality of user terminals that detect PII (Personal Identifiable Information) from input user questions and transfer user-neutral questions that convert the PII into neutral information, and a management server that receives the user-neutral questions and trains a management language model that generates common response patterns for each neutral question pattern.

Description

[Technical Field]

[0001] The present invention relates to an on-device based system and method for preventing personal information leakage and personalized response. [Background technology]

[0002] Recently, generative conversational AI such as ChatGPT has been attracting a lot of attention, but concerns about the leakage of personal identifiable information (PII) and the lack of personalization capabilities of AI have become apparent.

[0003] In the case of existing smartphones, only a small amount of PII, such as fingerprints and facial recognition information, is processed within the user's device without being sent to a management server. However, the amount of PII that needs to be stored is much larger in the case of the recently emerging generative dialogue AI, and models such as ChatGPT are emerging that can answer questions not only through text questions but also through images and videos, which contain much larger amounts of data. As users use personal devices for long periods of time, the amount of information accumulated increases further, and the existing method of managing only a small amount of PII on the device has its limitations.

[0004] On the other hand, personalization of conversational AI is essential to provide responses that match the user's tendencies, but there is a problem in that this process must be complemented with technology to prevent the leakage of PII.

[0005] Therefore, in order to solve the above problem, the present invention proposes an artificial intelligence agent that operates on a user terminal such as a smartphone, without transmitting PII to a management server such as a cloud. Summary of the Invention [Problem to be solved by the invention]

[0006] The present invention aims to solve the above-mentioned problems by providing an on-device-based system and method for preventing leakage of personal information and personalized responses, in which PII exists only on the user terminal and is not leaked to the outside, and personalized responses are generated through a program within the user terminal.

[0007] However, the technical problem that this embodiment aims to achieve is not limited to the above-mentioned technical problem, and other technical problems may exist. [Means for solving the problem]

[0008] As a technical means for solving the above-mentioned technical problems, an on-device-based personal information leakage prevention and personalized response system according to a first aspect of the present invention includes a plurality of user terminals that detect PII (Personal Identifiable Information) from input user questions and transmit user-neutral questions that convert the PII into neutral information, and a management server that receives the user-neutral questions and trains a management language model that generates a common response pattern for each neutral question pattern.

[0009] An on-device-based user terminal for preventing leakage of personal information and providing personalized responses according to a second aspect of the present invention includes a communication module, a memory storing a personalized response program, and a processor for executing the personalized response program, wherein the personalized response program detects PII from an input user question, converts the PII into neutral information, and transfers the user-neutral question converted into neutral information to a management server.

[0010] A personalized response method performed by an on-device-based personal information leakage prevention and personalized response system according to a third aspect of the present invention includes: (a) a step in which a management server receives user-neutral questions from a plurality of user terminals, in which PII contained in the user questions is converted into neutral information; and (b) a step in which the management server receives the user-neutral questions and trains a management language model that generates a common response pattern for each neutral question pattern. [Effects of the Invention]

[0011] According to any of the above-mentioned solutions to the problems of the present invention, personalized responses can be generated through a program within the device, without PII being leaked to the outside, and a personalized AI question and answer service can be provided. [Brief explanation of the drawings]

[0012] [Figure 1] 1 is a block diagram of an on-device-based personal information leakage prevention and personalized response system according to an embodiment of the present invention. [Figure 2] 1 is a diagram illustrating an on-device-based personal information leakage prevention and personalized response system according to an embodiment of the present invention. [Figure 3] FIG. 2 illustrates an example of a PII according to one embodiment of the present invention. [Figure 4] 10 is a block diagram of an on-device-based personal information leakage prevention and personalized response user terminal according to another embodiment of the present invention. [Figure 5] 10 is a flowchart illustrating a method for preventing leakage of personal information and providing a personalized response performed in a user terminal according to another embodiment of the present invention. [Figure 6] FIG. 10 is a diagram illustrating a personalized response program executed on a user terminal according to another embodiment of the present invention. [Figure 7] FIG. 10 is a diagram illustrating data interaction between a management server and a user terminal according to another embodiment of the present invention. [Figure 8] 10 is a flow chart illustrating a personalized response method performed in an on-device-based personal information leakage prevention and personalized response system according to another embodiment of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0013] Hereinafter, with reference to the accompanying drawings, embodiments of the present invention will be described in detail so that those skilled in the art can easily carry out the present invention. However, the present invention may be embodied in various different forms and is not limited to the embodiments described herein. In order to clearly explain the present invention in the drawings, parts that are not relevant to the description will be omitted, and similar parts will be designated by similar reference numerals throughout the specification.

[0014] Throughout this specification, a part being "connected" to another part includes not only a "directly connected" part but also an "electrically connected" part with another element sandwiched therebetween. Furthermore, unless otherwise specified, a part "including" a certain component does not exclude other components but means that it can further include other components.

[0015] As used herein, the term "module" includes both hardware-implemented units and software-implemented units. Furthermore, one unit may be implemented using two or more pieces of hardware, or two or more units may be implemented by a single piece of hardware. However, a "module" is not limited to software or hardware. A "module" may reside on an addressable storage medium or implement one or more processors. Thus, by way of example, a "module" includes components such as software components, object-oriented software components, class components, and task components, processes, functions, attributes, procedures, subroutines, program code segments, drivers, firmware, microcode, circuits, data, databases, data structures, tables, arrays, and variables. The components and functions provided within a "module" may be combined into fewer components and modules, or further separated into additional components and modules. Furthermore, a component and a "module" may be embodied to implement one or more CPUs within a device.

[0016] A network refers to a connection structure that enables information exchange between nodes such as terminals and servers, and includes a local area network (LAN), a wide area network (WAN), the Internet (WWW), wired and wireless data communication networks, telephone networks, wired and wireless television communication networks, etc. Examples of wireless data communication networks include, but are not limited to, 3G, 4G, 5G, 3GPP (3rd Generation Partnership Project), LTE (Long Term Evolution), WIMAX (World Interoperability for Microwave Access), Wi-Fi, Bluetooth (registered trademark), infrared communication, ultrasonic communication, visible light communication (VLC), and LiFi.

[0017] FIG. 1 is a configuration diagram of an on-device-based personal information leakage prevention and personalized response system according to one embodiment of the present invention, and FIG. 2 is a diagram illustrating an on-device-based personal information leakage prevention and personalized response system according to one embodiment of the present invention.

[0018] Referring to FIG. 1, a personalized response system 1 may include a plurality of user terminals 10 , a management server 20 and a database 30 .

[0019] 2, the user terminal 10 can detect PII (Personal Identifiable Information) from the input user question and transfer the user-neutral question, which converts the PII into neutral information, to the management server 20. In this case, PII refers to various information that can be used to identify a specific individual, such as a resident registration number, name, email address, or other information.

[0020] The management server 20 can receive user-neutral questions from a plurality of user terminals 10 and train a management language model 210 that generates a common response pattern for each neutral question pattern.

[0021] The database 30 may store or provide, as learning data, user-neutral question and response data generated by interactions between the management server 20 and the user terminal 10. As an example, the learning data managed by the database 30 may be used for federated learning, in which data stored in multiple locations cooperate with each other to learn a model without directly sharing the data.

[0022] The management server 20 may be implemented as a computer or a mobile terminal connectable to a network. Here, the computer may include, for example, a notebook computer, a desktop computer, a laptop computer, etc., and the mobile terminal may include, for example, any kind of handheld-based wireless communication device such as various smartphones, tablet PCs, smart watches, etc., which are wireless communication devices that ensure portability and mobility.

[0023] In addition, the management server 20 can provide a management language model learned based on user-neutral questions to the user terminal 10. In this case, the management server 20 can operate on a cloud computing service model such as SaaS (Software as a Service), PaaS (Platform as a Service), or IaaS (Infrastructure as a Service), or can be constructed in the form of a private cloud, a public cloud, or a hybrid cloud.

[0024] Specifically, the management server 20 includes a communication module, a memory, and a processor. The communication module provides a communication interface necessary for providing signals in the form of packet data to be transmitted and received to the user terminal 10 in conjunction with a communication network. Here, the communication module may be a device including hardware and software necessary for transmitting and receiving signals, such as control signals or data signals, with other network devices via wired or wireless connections.

[0025] The memory stores a personalized response program and temporarily or permanently stores data processed by the processor. The memory may include a volatile storage medium or a non-volatile storage medium, but the scope of the present invention is not limited thereto.

[0026] The memory may also store other programs, such as an operating system for processing and controlling the processor, and may also perform a function for temporarily storing input or output data.

[0027] The memory may include at least one type of storage medium: flash memory type, hard disk type, multimedia card micro type, card type memory (e.g., SD or XD memory), RAM, ROM.

[0028] The processor executes the personalized response program and provides functions for controlling the terminal hardware in response to the execution of the program, i.e., the processor can perform hardware control functions such as necessary file system, memory allocation, network, basic libraries, timers, device control (display, media, input device, 3D, etc.), and other utilities in response to the execution of the program.

[0029] Here, the term "processor" may include any type of device capable of processing data, such as a processor. Here, the term "processor" may refer to a data processing device built into hardware, having a physically structured circuit for performing a function expressed by a code or command contained in a program. Examples of such data processing devices built into hardware include a microprocessor, a central processing unit (CPU), a processor core, a multiprocessor, an application-specific integrated circuit (ASIC), a field programmable gate array (FPGA), etc., but the scope of the present invention is not limited thereto.

[0030] A personalized response program according to an embodiment of the present invention can input user-neutral questions received from multiple user terminals 10 into the supervised language model 210 to generate a common response pattern for each neutral question pattern. As an example, the supervised language model 210 can be updated through federated continual learning to derive common response patterns that are commonly useful for multiple users. Information on the update of the supervised language model 210 can then be transferred to each user terminal 10 and used to update the response generation model 132 of each user terminal 10.

[0031] For example, the supervised language model 210 may extract sentence structure elements, such as words, phrases, and clauses, from text input in the user's neutral question, classify the sentence structure elements into predetermined neutral question patterns, and match and learn common response patterns corresponding to the neutral question patterns. For example, the neutral question patterns may be classified for the same context based on a linguistic distribution including the frequency and type of words in the user's neutral question.

[0032] The supervised language model 210 can learn PII patterns for each neutral question pattern by utilizing the position of neutral information in the sentence structure for the user-neutral question. At this time, the neutral information is used to learn PII patterns for each neutral question pattern. <mask>This means that the text has been converted into a neutral word or a PII. <mask>By knowing the location of the token, the location of the neutral information can be known.

[0033] For example, each user terminal 10 can generate various patterns containing PII in user questions through dialogue. As an example, a first user can generate a pattern (first PII pattern) in which PII is included at position A in a neutral question pattern with a sentence structure of "ABC." As another example, a second user can generate a pattern (second PII pattern) in which PII is included at position C in a neutral question pattern with a sentence structure of "ABC." In other words, depending on the user's situation, even the same word may be determined to be PII for each user differently. For example, "Google" may be a simple individual word for a first user, but may be PII for a second user as workplace information.

[0034] Therefore, the management server 20 can grasp, for each neutral question pattern received from a plurality of user terminals 10, at what position PII was converted into neutral information and in what context.

[0035] FIG. 3 is a diagram illustrating an example of a PII according to one embodiment of the present invention.

[0036] Referring to FIG. 3, PII can be divided into direct identifiers and quasi-identifiers that can identify a specific individual. Exemplarily, a direct identifier refers to information in a PII dataset that can directly identify a specific individual by itself. For example, a direct identifier may include a name, a resident registration number, an address, a telephone number, an email address, etc. Furthermore, a quasi-identifier refers to information that can indirectly identify a specific individual in combination with other quasi-identifiers. For example, a quasi-identifier may include age, gender, political leanings, religion, habits, etc. A method for detecting PII will be described below with reference to FIGS. 5 to 7.

[0037] FIG. 4 is a configuration diagram of an on-device-based user terminal for preventing leakage of personal information and providing personalized responses according to another embodiment of the present invention; FIG. 5 is a flow chart illustrating a method for preventing leakage of personal information and providing personalized responses performed in a user terminal according to another embodiment of the present invention; FIG. 6 is a diagram illustrating a personalized response program executed in a user terminal according to another embodiment of the present invention; and FIG. 7 is a diagram illustrating data interaction between a management server and a user terminal according to another embodiment of the present invention.

[0038] Referring to FIG. 4, the user terminal 10 may include a communication module 110 , a memory 120 , a processor 130 and a database 140 .

[0039] The communication module 110 can receive information on an updated management language model from the management server 20 and transfer it to the processor 130. Here, the communication module 110 may be a device including hardware and software necessary to transmit and receive signals, such as control signals or data signals, with other network devices via wired or wireless connections.

[0040] The memory 120 may have a personalized response program recorded therein. The personalized response program detects PII from the input user query, converts the PII into neutral information, and transmits the converted user-neutral query to the management server. Here, the memory 120 may include a volatile storage device that requires power to maintain stored information, as well as a magnetic storage medium or a flash storage medium, although the scope of the present invention is not limited thereto.

[0041] The memory 120 may store a separate program, such as an operating system for processing and controlling the processor 130, and may also perform a function for temporarily storing input or output data.

[0042] The processor 130 executes a personalized response program (hereinafter, "program") stored in the memory 120, and by executing the program, provides a function for controlling the hardware of the user terminal 10. That is, by executing the program, the processor 130 can perform hardware control functions such as necessary file systems, memory allocation, networking, basic libraries, timers, device control (display, media, input devices, 3D, etc.), and other utilities.

[0043] Referring to FIG. 5, the processor 130 detects PII from the input user question (S110), converts the PII into neutral information (S120), and transfers the user neutral question converted into neutral information to the management server 20 (S130).

[0044] The processor 130 may include any type of device capable of processing data. For example, the processor 130 may refer to a data processing device built into hardware, having a physically structured circuit for performing functions expressed by code or commands contained in a program. Examples of such data processing devices built into hardware include a microprocessor, a central processing unit (CPU), a processor core, a multiprocessor, an application-specific integrated circuit (ASIC), a field programmable gate array (FPGA), etc., but the scope of the present invention is not limited thereto.

[0045] The database 140 stores or provides data required for the user terminal 10 under the control of the processor 130. Illustratively, the database 140 may include a PII database 141. Referring to FIG. 6, the PII database 141 may store PII extracted from user queries. The database 140 may be included as a separate component from the memory 120 or may be constructed in a partial area of ​​the memory 120.

[0046] 6, the processor 130 may include a detailed module that performs various functions in response to the execution of the personalized response program, where the detailed module may include a PII detection model 131 and a response generation model 132.

[0047] 7, when the user terminal 10 receives update information for the administration language model 210 from the administration server 20, it can update the response generation model 132 or the PII detection model 131. As an example, the PII detection model 131 and the response generation model 132 can perform continual learning so that the already trained models can be updated with new data sets, environments, etc. while maintaining their existing performance.

[0048] For example, the PII detection model 131 may utilize an existing language recognition model and may be configured with an encoder structure. Similarly, the response generation model 132 may utilize an existing language generation model and may be configured with a decoder or encoder-decoder structure. For example, each model may employ various elements such as a Transformer model, a convolutional neural network (CNN), or a recurrent neural network (RNN) as its basic building blocks.

[0049] 7, the PII detection model 131 can detect PII from an input user question, convert the PII into neutral information, and transfer the user-neutral question converted into neutral information to the management server 20. At this time, the PII detection model 131 can mask the PII extracted from the user question or convert it into predetermined words to generate a user-neutral question in which the PII has been filtered.

[0050] As an example, when the PII detection model 131 converts PII into neutral information, it can convert the PII into the corresponding text by specifying a representative neutral word for each type of individual classified as PII from the user question. For example, the question "My name is Yoo Sang-won. What is your name?" can be converted into a neutral question "My name is Hong Gil-dong. What is your name?". As another example, when the PII detection model 131 converts PII into neutral information, it can convert the PII extracted from the user question into neutral information. <mask>You can transform queries with simple masking using tokens. For example, the question "My name is Yoo Sang-won. What is your name?" can be transformed into "My name is <mask>It can be transformed into a neutral question: "What is your name?"

[0051] In this way, the management server 20 can train the management language model 210 using the user-neutral question data from which PII has been filtered, and provide a common response for each neutral question pattern.

[0052] As another example, the management server 20 can learn the PII pattern of each user while learning a plurality of user-neutral question patterns. Here, the PII pattern is, as described above, a neutral word in a sentence structure or <mask>The position of the token can be used to learn each neutral query pattern. In one embodiment, the management server 20 can provide the PII pattern of the second user terminal to the first user terminal as update information. For example, in a first user query "ABC," the first user typically includes PII at position A, but the second user includes PII at position C. Therefore, information indicating that there is a possibility that PII may be included at both positions A and C in the query structure can be provided to the first user terminal as update information. The first user terminal can then input a first user query similar to the PII pattern of the second user terminal, rather than its original format. In this case, the updated PII detection model 131 of the first user terminal can easily recognize PII from the first user query. This allows each user terminal 10 to improve the PII filtering function of the PII detection model 131. Furthermore, because the pattern including PII is updated rather than updating the PII of other users, the leakage of personal information can be prevented.

[0053] For example, the PII detection model 131 may determine whether characters or words constituting a user question are personal information based on the initial PII stored in the PII database 141, extract PII from the user question based on named entity recognition (NER), and store the PII in the PII database 141. For example, the process of constructing the PII database 141 must be performed in advance, and the PII database 141 may be constructed by extracting some of the entities classified as common PII from the user question as initial PII through a text recognition methodology such as named entity recognition (NER), or by a user directly registering the initial PII. Furthermore, the PII detection model 131 may determine whether each entity constituting the text of a newly input user question is PII based on the PII database 141, thereby expanding the PII database 141 and improving the accuracy of PII detection.

[0054] For example, the PII detection model 131 can learn in a continuous learning manner based on the PII database 141, which is updated based on user questions and usage behavior, to identify PII patterns. Therefore, the PII detection model 131 can determine whether a user question contains PII and transfer user-neutral questions, in which the PII has been removed, to the management server 20. That is, by using the PII database 141 as training data for the PII detection model 131, the PII detection model 131 can also detect text entities related to the PII database 141, thereby enhancing the generalization ability of PII detection. The loss function applied here may be a type of loss function commonly used in training language encoder models. For example, a negative log likelihood (NLL) function may be applied as a loss function in a masked language modeling (MLM) training format.

[0055] The response generation model 132 can understand the context of the user's question through natural language processing analysis and generate a response. The response generation model 132 can also be configured as a language model that learns user question patterns using user questions. Here, the user question patterns may be learned for each same context based on a language distribution including the frequency and type of words in the user's question.

[0056] For example, the response generation model 132 may use a collection of user questions in the form of natural language text as training data. Here, the training data may be steadily updated as user question data input by users is accumulated. This allows the response generation model 132 to learn user question patterns based on the accumulated user questions, and each user terminal 10 may provide users with personalized, natural responses.

[0057] For example, the learning method of the response generation model 132 may be such that the distribution of languages ​​generated by the initial model approaches the distribution of query languages ​​input by the user. A loss function that measures the distance between distributions may be applied to learning so that the text distribution of data generated by the response generation model 132 approaches the text distribution composed of user question data. To prevent catastrophic forgetting while continuously learning steadily updated user question data, a loss function that reduces the distance between the distribution generated by the model before learning and the distribution generated by the model after learning may be additionally applied to learning each time data is learned. For example, cross entropy may be applied as a loss function that measures the difference between distributions.

[0058] In another embodiment, in the on-device-based personal information leakage prevention and personalized response system of the present invention, if the service operators of the management server 20 and the user terminal 10 are the same or cooperate with each other, the management server 20 can perform federated learning of the management language model 210 and provide update information for each model to the user terminal 10. In other words, the management server 20 can exchange learning information with each user terminal 10 and interact with each other.

[0059] In a further embodiment, if the service operators of the management server 20 and the user terminal 10 are different and do not have mutual access rights, the user terminal 10 can only transfer a neutral question to the management server 20 and receive a common response to the neutral question. That is, each user terminal 10 can only provide a personalized response service using the user question data and the learned response generation model 132. In this case, the management server 20, as a third party, can only perform the role of additionally providing a common response when the user terminal 10 generates a response through the management language model 210.

[0060] In the following, the description of the same components as those shown in the above-mentioned FIGS. 1 to 7 will be omitted.

[0061] FIG. 8 is a flow chart illustrating a method for providing a personalized response performed in an on-device-based personal information leakage prevention and personalized response system according to another embodiment of the present invention.

[0062] Referring to FIG. 8, the method for preventing leakage of personal information and personalized response performed by the on-device-based personalized response system 1 includes a step (S210) in which the management server 20 receives user-neutral questions from multiple user terminals 10, in which PII (Personal Identifiable Information) contained in the user questions is converted into neutral information, and a step (S220) in which the management server 20 receives the user-neutral questions and trains a management language model 210 that generates common response patterns for each neutral question pattern.

[0063] Here, the supervised language model 210 may be configured to learn PII patterns for each neutral question pattern using the location of neutral information in the sentence structure for the user-neutral question.

[0064] A personalized response method according to an embodiment of the present invention may also be embodied in the form of a recording medium containing computer-executable commands, such as program modules executed by a computer. A computer-readable medium may be any available medium accessible by a computer, including both volatile and nonvolatile media, and both removable and non-removable media. A computer-readable medium may also include computer storage media. A computer storage medium includes both volatile and nonvolatile, removable and non-removable media embodied in any method or technology for storing information, such as computer-readable commands, data structures, program modules, or other data.

[0065] Although the apparatus and methods of the present invention have been described with reference to specific embodiments, some or all of their components or operations may be implemented using a computer system having a general-purpose hardware architecture.

[0066] The above description of the present invention is for illustrative purposes only, and those skilled in the art will understand that the present invention can be easily modified into other specific forms without changing the technical spirit or essential characteristics of the present invention. Therefore, the above-described embodiments should be understood to be illustrative in all respects and not restrictive. For example, each component described as a single type can be implemented in a distributed form, and similarly, each component described as a distributed type can be implemented in a combined form.

[0067] The scope of the present invention is indicated by the claims that follow rather than by the above detailed description, and all modifications and variations that fall within the meaning and scope of the claims and their equivalents should be construed as being included within the scope of the present invention.< / mask> < / mask> < / mask> < / mask> < / mask>

Claims

1. On-device based personal information leakage prevention and personalized response system a plurality of user terminals that detect PII (Personal Identifiable Information) from an input user question and transfer user-neutral questions in which the PII is converted into neutral information; and The personalized response system includes a management server that receives the user-neutral questions and trains a management language model that generates a common response pattern for each neutral question pattern.

2. The user terminal A response generation model that understands the context of the user question through natural language processing analysis and generates a response; and The personalized response system of claim 1 , further comprising a PII detection model that masks or converts the PII extracted from the user query into predetermined words to generate the user-neutral query.

3. the response generation model is a language model that learns a user question pattern using the user question; The personalized response system of claim 2 , wherein the user question patterns are learned for each same context based on a linguistic distribution including frequency and type of words in the user questions.

4. The PII detection model is Based on the initial PII stored in advance in the PII database, determine whether the individual characters or words constituting the user query are personal information; 3. The personalized response system of claim 2, wherein the PII is extracted from the user query based on named entity recognition (NER) and stored in the PII database.

5. The personalized response system of claim 1 , wherein the supervised language model learns PII patterns for each neutral question pattern using the position of the neutral information in a sentence structure for the user-neutral question.

6. The personalized response system of claim 1 , wherein the PII is divided into direct identifiers and quasi-identifiers that can identify a specific individual.

7. In a user terminal for preventing leakage of personal information and personalized responses based on an on-device basis, communication module, a memory having a personalized response program stored therein; and a processor that executes the personalized response program; The personalized response program detects PII from the input user question, converts the PII into neutral information, and transfers the user-neutral question converted into neutral information to a management server.

8. The personalized response program A response generation model that understands the context of the user question through natural language processing analysis and generates a response; and The user terminal of claim 7 , further comprising a PII detection model that generates the user-neutral query by masking or converting the PII extracted from the user query into predetermined words.

9. the response generation model is a language model that learns a user question pattern using the user question; The user terminal according to claim 8 , wherein the user question pattern is learned for each same context based on a language distribution including frequency and type of words in the user questions.

10. The PII detection model is Based on the initial PII stored in advance in the PII database, determine whether the individual characters or words constituting the user query are personal information; 9. The user terminal of claim 8, wherein the PII is extracted from the user query based on named entity recognition (NER) and stored in the PII database.

11. In a method for preventing leakage of personal information and providing a personalized response in an on-device-based personalized response system, (a) a step in which a management server receives, from a plurality of user terminals, user-neutral queries in which PII (Personal Identifiable Information) included in the user queries has been converted into neutral information; and (b) a step of a management server receiving the user-neutral questions and training a management language model that generates common response patterns for each neutral question pattern;

12. The personalized response method of claim 11 , wherein the supervised language model learns a PII pattern for each neutral question pattern by utilizing the position of the neutral information in a sentence structure for the user-neutral question.

Images