Knowledge graph-based content search method and electronic device for performing same

Abstract

Provided is a method by which an electronic device provides a content search. The method comprises the steps of: providing a knowledge graph-based search result on the basis of a search word input during a search session, wherein the search session includes sequential search queries, and a corresponding search result is provided whenever each search query occurs; terminating the search session and identifying selected content as target content when the content is selected in the search results; analyzing, among the search queries, search results corresponding to previous search queries for which the target content has not been selected in the search session; and optimizing the knowledge graph-based search by adding a hidden connection line in the knowledge graph on the basis of the analysis result in order to provide, as a main item, the target content in the search results corresponding to the previous search queries.

Description

Knowledge graph-based content search method and electronic device performing the method

[0001] The present disclosure relates to a method for providing content search based on a knowledge graph, generating and updating a knowledge graph for content search optimization, and an electronic device for performing content search.

[0002] As the storage capacity of electronic devices increases and the volume of stored content grows massive, efficiently searching for desired content within the device is becoming increasingly difficult. Generally, electronic devices search for content based on user search input, but this presents a problem where search results appear either excessively vast or too narrow. Although search methods utilizing knowledge graphs have been introduced to address this, they have limitations, as it is difficult to obtain desired results if the user fails to select appropriate search terms. Therefore, there is a need for effective search optimization methods that can enhance the user experience without distorting search results.

[0003] According to one aspect of the present disclosure, a method may be provided in which an electronic device provides content search. The method may include the step of providing knowledge graph-based search results during a search session based on search term input. The search session may include sequential search queries, and a corresponding search result may be provided whenever each search query occurs. The method may include the step of terminating the search session and identifying the selected content as target content when content is selected from the search results. The method may include the step of analyzing search results corresponding to previous search queries in which the target content was not selected among the search queries within the search session. The method may include the step of optimizing the knowledge graph-based search by adding hidden connections within the knowledge graph based on the analysis results to provide the target content as a main item in the search results corresponding to the previous search queries.

[0004] According to one aspect of the present disclosure, an electronic device for providing content search may be provided. The electronic device may include at least one processor, a memory for storing instructions, and a display. By executing the instructions by the at least one processor, the electronic device may provide knowledge graph-based search results during a search session based on search term input. The search session may include sequential search queries, and a corresponding search result may be provided whenever each search query occurs. By executing the instructions by the at least one processor, the electronic device may terminate the search session and identify the selected content as target content when content is selected from the search results. By executing the instructions by the at least one processor, the electronic device may analyze search results corresponding to previous search queries within the search session in which the target content was not selected. By executing the above instructions by the at least one processor, the electronic device can optimize the knowledge graph-based search by adding hidden connections within the knowledge graph based on the analysis results to provide the target content as a main item in the search results corresponding to the previous search queries.

[0005] According to one aspect of the present disclosure, a computer-readable recording medium may be provided having a program recorded thereon for executing any one of the methods described above and below, wherein an electronic device searches for content based on a knowledge graph and performs optimization of the knowledge graph-based search.

[0006] FIG. 1 is a diagram illustrating an example of a knowledge graph-based search provided by an electronic device according to one embodiment of the present disclosure.

[0007] FIG. 2 is a block diagram illustrating the configuration of an electronic device according to one embodiment of the present disclosure.

[0008] FIG. 3 is a flowchart illustrating the operation of an electronic device according to one embodiment of the present disclosure optimizing a knowledge graph-based search.

[0009] FIG. 4 is a diagram illustrating a knowledge graph used by an electronic device according to one embodiment of the present disclosure when performing a search.

[0010] FIG. 5 is a diagram illustrating a search process performed in an electronic device according to one embodiment of the present disclosure.

[0011] FIG. 6 is a diagram illustrating the operation of an electronic device analyzing search results according to one embodiment of the present disclosure.

[0012] FIG. 7 is a diagram illustrating the operation of an electronic device according to one embodiment of the present disclosure to perform a search using a knowledge graph including hidden connection lines.

[0013] FIG. 8 is a diagram illustrating the operation of adding a hidden connection line to a knowledge graph using an electronic device according to one embodiment of the present disclosure.

[0014] FIG. 9 is a diagram illustrating the operation of adding a hidden connection line to a knowledge graph using an electronic device according to one embodiment of the present disclosure.

[0015] FIG. 10 is a diagram illustrating the operation of an electronic device according to one embodiment of the present disclosure providing search results using a knowledge graph including hidden connection lines.

[0016] FIG. 11 is a diagram illustrating the operation of adding a hidden connection line to a knowledge graph using an electronic device according to one embodiment of the present disclosure.

[0017] FIG. 12 is a diagram illustrating the operation of an electronic device according to one embodiment of the present disclosure to generate a knowledge graph with hidden connection lines added corresponding to a search session.

[0018] FIG. 13 is a diagram illustrating the operation of an electronic device according to one embodiment of the present disclosure to build a knowledge graph-based personalized database.

[0019] FIG. 14 is a drawing for exemplarily illustrating the use of a knowledge graph generated in an electronic device according to one embodiment of the present disclosure.

[0020] FIG. 15 is a drawing for illustrating an example of a knowledge graph generated in an electronic device according to one embodiment of the present disclosure.

[0021] FIG. 16 is a flowchart illustrating the operation of an electronic device generating a knowledge graph according to one embodiment of the present disclosure.

[0022] FIG. 17 is a diagram illustrating a first data group and a second data group clustered from user data related to a plurality of applications in an electronic device according to one embodiment of the present disclosure.

[0023] FIG. 18 is a flowchart illustrating the process of expanding and acquiring a second data group in an electronic device according to one embodiment of the present disclosure.

[0024] FIG. 19 is a drawing for illustrating the expansion of a second data group in an electronic device according to one embodiment of the present disclosure.

[0025] FIG. 20 is a flowchart illustrating the process of obtaining a knowledge graph based on interrelated user data in an electronic device according to one embodiment of the present disclosure.

[0026] The terms used in this specification will be briefly explained, and the present disclosure will be described in detail. In the present disclosure, the expression "at least one of a, b, or c" may refer to "a," "b," "c," "a and b," "a and c," "b and c," "all of a, b, and c," or variations thereof.

[0027] The terms used in this disclosure have been selected to be as widely used and general as possible, taking into account their functions within this disclosure; however, these terms may vary depending on the intent of those skilled in the art, case law, the emergence of new technologies, etc. Additionally, in specific cases, terms have been selected at the applicant's discretion, and in such cases, their meanings will be described in detail in the relevant explanatory sections. Therefore, terms used in this disclosure should be defined not merely by their names, but based on their meanings and the overall content of this disclosure.

[0028] Singular expressions may include plural expressions unless the context clearly indicates otherwise. Terms used herein, including technical or scientific terms, may have the same meaning as generally understood by those skilled in the art as described in this specification. Additionally, terms including ordinal numbers, such as "first" or "second," used in this specification may be used to describe various components, but said components should not be limited by said terms. Such terms are used solely for the purpose of distinguishing one component from another.

[0029] When a part of a specification is described as "comprising" a certain component, this means that, unless specifically stated otherwise, it does not exclude other components but may include additional components. Furthermore, terms such as "part" or "module" as used in the specification refer to a unit that processes at least one function or operation, and this may be implemented in hardware or software, or as a combination of hardware and software.

[0030] In the present disclosure, a "knowledge graph" refers to a method for managing and searching knowledge information, meaning a knowledge base in graph form based on a knowledge base that stores knowledge information and a graph that can be analyzed as a network structure. A knowledge graph is a graph model that implements knowledge accumulated in a knowledge base through node and edge relationships. A knowledge graph can be used to integrate data using graph data models, topologies, etc. In order for knowledge to be interconnected and integrated using a knowledge graph, a schema is implemented through an ontology, and a structure and dictionary (terms) that can be shared are used.

[0031] Semantic information containing common sense and fact knowledge is organized into connections between nodes and edges, and by referencing this, various types of data can be transformed into the form of a knowledge graph. Methods for representing a knowledge graph may include, but are not limited to, Labeled Property Graph (LPG), where nodes and edges can each have attributes, and Resource Description Framework (RDF), which expresses relationships using a triple-fact structure of subject-predicate-object. A knowledge graph can be generated by recognizing entities from various data, connecting them to appropriate entities in an existing knowledge base, and extracting relationships between entities.

[0032] Knowledge graphs can be utilized to improve the performance of artificial intelligence. Knowledge graphs can be used in models such as Graph Neural Networks (GNNs) and Graph Convolutional Neural Networks (GCNs), and can be used to provide explanations for results in Explainable Artificial Intelligence (XAI). Below, with reference to the attached drawings, embodiments of the present disclosure are described in detail so that those skilled in the art can easily implement the present invention. However, the present disclosure may be embodied in various different forms and is not limited to the embodiments described herein. Furthermore, parts unrelated to the description have been omitted from the drawings to clearly explain the present disclosure. Additionally, throughout the specification, different reference numbers have been used for the same components for convenience of explanation.

[0033] The present disclosure will be described below with reference to the attached drawings.

[0034] FIG. 1 is a diagram illustrating an example of a knowledge graph-based search provided by an electronic device according to one embodiment of the present disclosure.

[0035] In one embodiment, the electronic device (100) may provide a search function to a user. The electronic device (100) may search for one or more contents (30) based on a knowledge graph (20) based on user input entering a search term into a search interface (10).

[0036] The knowledge graph (20) may refer to a personal knowledge graph that standardizes information related to an individual user by expanding the knowledge graph using acquired user data, based on a general knowledge graph constructed in a standardized form based on common sense or factual knowledge.

[0037] The electronic device (100) can retrieve one or more contents (30) stored in the storage of the electronic device (100) based on a knowledge graph based on user input. For example, the electronic device (100) can retrieve photos, videos, audio, messages, emails, calendar events, contacts, memos, etc.

[0038] Meanwhile, when a user searches for one or more contents (30) on an electronic device (100), the user may not easily find the desired result (e.g., a specific photo, video, etc.). For example, the user of the electronic device (100) may not find the desired content until they enter a search term N-1 times, and may find and select the desired content from the results of entering the Nth search term. That is, from the user's perspective, they have entered a search term to search for content based on their memory, but there may be cases where the search algorithm in the electronic device (100) fails to search for and provide content using that search term. In this disclosure, the search process through the user's input of a series of search terms is referred to as a search session.

[0039] In one embodiment, the electronic device (100) can optimize the knowledge graph (20) to improve the user's search experience. The electronic device (100) can analyze search results corresponding to a search term based on whether the search failed or succeeded, and update the knowledge graph (20) based on the analysis results. The electronic device (100) can add connection lines between nodes within the knowledge graph and enable / disable the added connection lines to allow the user to receive the desired search results. When the knowledge graph (20) is optimized, when the user performs the same or similar search in the future, the electronic device (100) can provide the user with the desired search results for the search terms within the search session early. For example, in the example described above, the user succeeded in the search through the Nth search, but if the user uses the same set of search terms in a subsequent search, the electronic device (100) can provide the corresponding content with fewer than N search term inputs using the updated knowledge graph (20).

[0040] The electronic device (100) of the present disclosure may be implemented as various types of devices that provide a content search function. For example, the electronic device may include devices capable of collecting data related to a user, storing and managing content, and executing a content search function, such as a smart TV including a display, a smartphone, a tablet PC, a laptop PC, and smart signage, but is not limited thereto.

[0041] Specific operations for an electronic device to search for one or more contents (30) based on a knowledge graph (20) will be described in more detail through the drawings and descriptions therefor.

[0042] FIG. 2 is a block diagram illustrating the configuration of an electronic device according to one embodiment of the present disclosure.

[0043] Referring to FIG. 2, an electronic device (100) according to one embodiment may include a processor (110), a memory (120), and a display (130).

[0044] The processor (110) can control the overall operations of the electronic device (100). By executing one or more instructions of a program stored in the memory (120) of the electronic device (100), the processor (110) can control the operations of the electronic device (100) to search for content based on a knowledge graph and provide search results. For example, the processor (110) can handle various tasks including processing arithmetic and logical operations of processes running in the electronic device (100), managing the main memory (122), and transferring data to storage (124).

[0045] The processor (110) may include a processing circuit. The processing circuit may include an operation unit that performs arithmetic and logical operations, a control unit that interprets commands and controls execution, a register that stores data, and a cache memory, but is not limited thereto.

[0046] The processor (110) may be composed of at least one of, for example, a Central Processing Unit (CPU), a Microprocessor, a Graphic Processing Unit (GPU), ASICs (Application Specific Integrated Circuits), DSPs (Digital Signal Processors), DSPDs (Digital Signal Processing Devices), PLDs (Programmable Logic Devices), FPGAs (Field Programmable Gate Arrays), an Application Processor (AP), a Neural Processing Unit (NPU), or an AI-dedicated processor designed with a hardware structure specialized for processing AI models, but is not limited thereto.

[0047] In one embodiment, there may be one or more processors (110). If there is one or more processors (110), the operations of the present disclosure may be performed by one or more processors by executing instructions and / or programs stored in memory (120) individually or collectively. If the method according to one embodiment of the present disclosure includes a plurality of operations, the plurality of operations may be performed by one processor (110) or by a plurality of processors (110).

[0048] For example, when a first operation, a second operation, and a third operation are performed in an electronic device (100) by a method according to one embodiment, the first operation, the second operation, and the third operation may all be performed by a first processor, and some of the first to third operations may be performed by a first processor for processing general-purpose tasks and the remaining operations may be performed by a second processor for processing second machine learning-related tasks. However, the embodiments of the present disclosure are not limited thereto.

[0049] One or more processors (110) according to the present disclosure may be implemented as a single-core processor or as a multi-core processor. When a method according to one embodiment of the present disclosure includes a plurality of operations, the plurality of operations may be performed by a single core or by a plurality of cores included in one or more processors (110).

[0050] The detailed operations of the electronic device (100) performed by the processor (110) will be described in detail with reference to the drawings below.

[0051] The memory (120) can store data processed by the electronic device (100).

[0052] The memory (120) may include a main memory (122) that stores data currently being processed in the electronic device (100). The main memory (122) may store programs and data currently being executed by the processor (110) to allow the processor (110) to access the data quickly. The main memory (122) may include volatile memory such as, for example, RAM (Random Access Memory) or SRAM (Static Random Access Memory), but is not limited thereto.

[0053] The memory (120) may include storage (124) for permanently storing large amounts of data (e.g., programs, applications, system files, media files, etc.). The storage (124) may include non-volatile memory including at least one of, for example, a hard disk drive (HDD), a solid-state drive (SSD), an optical drive (e.g., a CD), a flash drive, a ROM (Read-Only Memory), an EEPROM (Electrically Erasable Programmable Read-Only Memory), and a PROM (Programmable Read-Only Memory), but is not limited thereto.

[0054] The display (130) can output a video signal to the screen of the electronic device (100) under the control of the processor (110). For example, the display (130) can display a graphic interface to provide a search function to the user and can output a video signal to the screen to provide one or more contents representing search results to the user as the search is performed.

[0055] FIG. 3 is a flowchart illustrating the operation of an electronic device according to one embodiment of the present disclosure optimizing a knowledge graph-based search.

[0056] In operation 310, the electronic device (100) can provide search results based on a knowledge graph during a search session based on search term input.

[0057] In one embodiment, the electronic device (100) may provide a content search function. The electronic device (100) may search for content stored in the storage of the electronic device (100) based on user input. For example, the electronic device (100) may search for photos, videos, audio, messages, emails, calendar events, contacts, memos, etc.

[0058] Search term input can be obtained from a user. User input entering a search term can be text input or voice-based input. If the user input is voice input, the electronic device (100) can convert the voice into text using speech recognition technology (e.g., Speech-to-Text).

[0059] A search session may refer to a series of search activities in which a user utilizes the search function of the electronic device (100) to find content within the electronic device (100). A search session may include sequential search queries. During a search session, when a user enters a search term into the electronic device (100), a search query is generated, and the electronic device (100) may repeat the process of providing search results to the user based on the search query and the knowledge graph.

[0060] A user of the electronic device (100) may repeat entering a search term until a desired search result is obtained in order to find specific content. For example, if the user enters search term A, search result A may be provided. Subsequently, if the user enters search term B, search result B may be provided. In one embodiment, the electronic device (100) may provide a multiple search term input function. For example, if the user enters two or more search terms, a search query corresponding to two or more search terms is generated, and the electronic device (100) may perform a search based on the generated search query and the knowledge graph.

[0061] In operation 320, the electronic device (100) can terminate the search session when content is selected from the search results and identify the selected content as target content.

[0062] In one embodiment, the electronic device (100) can determine whether a search is successful for search queries included in a search session. For example, when search results are provided to a user based on a search query, if the user does not select content within the search results, this may mean that the electronic device (100) failed to provide the user with the content desired by the user as a search result. Therefore, the electronic device (100) may determine that the case where the user does not select content from the search results is a search failure. Alternatively, if the user selects content within the search results, this may mean that the electronic device (100) provided the user with a search result containing the desired content. Therefore, the electronic device (100) may determine that the case where the user selects content from the search results is a search success.

[0063] When content is selected from the search results, the electronic device (100) can determine that the selected content is the target content. The target content may refer to content that corresponds to the search intent that the user of the electronic device (100) intended to find through content search. When content is selected, the electronic device (100) can identify a node connected to the target content within the knowledge graph to obtain target content information and node information.

[0064] When the electronic device (100) determines that the search is successful by selecting target content, it can terminate the search session and obtain information about the search terms included in the search session and the search queries corresponding to the search terms.

[0065] In operation 330, the electronic device (100) can analyze search results corresponding to previous search queries in which target content was not selected among the search queries within the search session.

[0066] Previous search queries within a search session in which target content was not selected can be classified as search failures because, even if search results are displayed, the user's content selection was not made. When the search session of the electronic device (100) ends, the search results for the search queries determined to have failed among the search queries included in the search session can be analyzed.

[0067] The electronic device (100) can analyze the search results for search queries that failed to search and analyze the cause of the search failure. The cause of the search failure may include, for example, a case where the search results provided based on the search query contained target content that the user was looking for, but the user could not select the target content because there were too many search results (in the case of a broad search query). Or, for example, a case where the search results provided based on the search query did not contain target content (in the case of a narrow search query). The electronic device (100) can classify the search queries according to the cause of the search failure.

[0068] In operation 340, the electronic device (100) can optimize knowledge graph-based search by adding hidden connections within the knowledge graph based on the analysis results.

[0069] In one embodiment, the electronic device (100) may analyze search results to provide target content as a main item in search results corresponding to previous search queries. The electronic device (100) may analyze the search results of previous search queries to identify the cause of search failure and identify knowledge graph optimization conditions corresponding to the classification of the search query that failed the search. For example, if the search query is classified as a broad search query, the electronic device (100) may add hidden connections to the knowledge graph according to defined conditions that allow the target content to be included in the search results while narrowing the search range. Or, for example, if the search query is classified as a narrow search query, the electronic device (100) may add hidden connections to the knowledge graph according to defined conditions that allow the target content to be included in the search results while maintaining or expanding the search range. The defined conditions for the electronic device (100) to add hidden connections to the knowledge graph will be described in FIGS. 8, FIGS. 9, and FIGS. 10.

[0070] In one embodiment, a hidden connection line added to a knowledge graph can provide a new search path by connecting a set of nodes within the knowledge graph. The hidden connection line is set to a disabled state and can be enabled only when a search term included in the search session is entered when a user uses the search function at a later time. The electronic device (100) can enable the hidden connection line of the knowledge graph based on the entry of a search term included in the search session and provide a search result based on the knowledge graph including the enabled hidden connection line.

[0071] FIG. 4 is a diagram illustrating a knowledge graph used by an electronic device according to one embodiment of the present disclosure when performing a search.

[0072] In the present disclosure, a knowledge graph may refer to a database capable of structurally representing entities and relationships between entities and supporting semantic searching for managing and exploring various types of content such as text, images, and videos.

[0073] In one embodiment, the electronic device (100) may collect user data, including user activity, time of event occurrence, location, weather, etc., from applications running on the electronic device (100) or from one or more sensors mounted on the electronic device (100). The electronic device (100) may generate activity information based on the user data, generate entities representing the activity information and connect them to each other, and generate a knowledge graph by storing attribute information of each entity. The electronic device (100) may construct a knowledge graph by defining entities and the relationships between entities based on the collected user data. For example, the electronic device (100) may define entities and the relationships between entities, such as 'travel - activity - restaurant', based on the collected data.

[0074] A knowledge graph may include nodes representing entities related to the user of the electronic device (100) and connecting lines between nodes representing the relationships between the entities. The connecting lines may also be referred to as edges. The entities of the knowledge graph may be semantic concepts related to the user's activity information and may be represented as nodes within the knowledge graph. For example, concepts related to the user's activities, such as travel, business trips, and hobbies (e.g., taking photos), may be defined as entities. In this case, the nodes may be detailed data elements representing the attributes of the entities. For example, if the entity is travel, specific data elements such as overseas travel, domestic travel, family travel, and travel destinations (e.g., Seoul) may be represented as nodes in the knowledge graph.

[0075] In one embodiment, the electronic device (100) can search for content based on a knowledge graph and provide search results. For example, the electronic device (100) can receive a search term (410) from a user, convert it into a search query, and search the knowledge graph based on the search query to retrieve search results. The electronic device (100) may use an Internationalized Resource Identifier (IRI) to identify each entity and the relationships between entities for the knowledge graph search. The search results may include various types of content, such as text, images, and videos.

[0076] As a content search function is provided in the electronic device (100), the user can find the content they want by entering a search term (410) into the electronic device (100). In the present disclosure, content corresponding to the user's search intent will be referred to as target content (420). When a search term (410) is entered, the electronic device (100) can convert the search term into a search query and start searching along a search path based on a knowledge graph.

[0077] For example, the electronic device (100) can identify a starting node corresponding to a search query and search for nodes within a predetermined range from the starting node. The predetermined range may mean that the number of hops, which is the unit for skipping connecting lines within a knowledge graph, is less than or equal to a predetermined number (e.g., N-hop search), but is not limited thereto.

[0078] When a user uses the search function of an electronic device (100), the user writes a search term (410) to find the target content (420) by relying on their memory, so the search term (410) may be a word (or sentence) that is not suitable for finding the target content (420). For example, the search term (410) may cause a wide range of search results so that the user cannot find the target content (420) in the search results, or the search term (410) may cause a narrow range of search results so that the target content (420) is not included in the search results. In addition, if the desired search results do not appear, the user may repeatedly enter the search term until the target content (420) is found.

[0079] In one embodiment, the electronic device (100) may add hidden connections (430) to the knowledge graph so that the search results corresponding to the search term (410) are not too broad, while the search results include target content (420). The hidden connections (430) may connect sets of nodes within the knowledge graph to provide new search paths.

[0080] The electronic device (100) can create a hidden connection line (430) corresponding to a specific search session in the knowledge graph to improve the search experience for subsequent searches using the user's search terms within that search session. The hidden connection line (430) is set to a deactivated state and can be activated when a search term within that search session is entered. The hidden connection line (430) can remain deactivated when a search term other than the one within that search session is entered, thereby allowing search results for search terms outside that search session to remain unchanged. The electronic device (100) can operate the search algorithm at a low maintenance cost by creating and managing a hidden connection line (430) in the knowledge graph, rather than creating, storing, and managing synonyms or related terms for a large number of search terms. The operations of the electronic device (100) optimizing knowledge graph-based search by adding a hidden connection line (430) within the knowledge graph will be described in more detail with reference to the drawings below.

[0081] In the example of Figure 4, for the sake of brevity, an example is shown where content is connected to only some nodes, but there may be multiple corresponding contents for each node in the knowledge graph. Specifically, if a specific node in the knowledge graph represents the user's activity information 'travel', photos related to that travel may be connected to that node.

[0082] In one embodiment, there may be multiple knowledge graphs. In this case, each of the multiple knowledge graphs may have a different title or topic and may be configured to include information corresponding to the set title or topic.

[0083] FIG. 5 is a diagram illustrating a search process performed in an electronic device according to one embodiment of the present disclosure.

[0084] In describing Fig. 5, an example is given in which a user searches for content using an electronic device (100), and a search attempt is repeated through the input of a search term until a target content corresponding to the user's search intention is found.

[0085] In one embodiment, the electronic device (100) can provide search results based on a knowledge graph during a search session.

[0086] For example, when search term A is entered, the electronic device (100) can convert search term A into a search query (search query A) and provide search result A based on a knowledge graph. If the user cannot find the target content in search result A, the user can enter the next search term, search term B, and receive search result B. In the same way, the user enters search terms sequentially, and the electronic device (100) can provide a corresponding search result whenever a sequential search query occurs.

[0087] A user of the electronic device (100) may select the target content if the desired content is included in the search results. When the electronic device (100) identifies that content is selected in the search results, it may determine that the selected content is the target content corresponding to the user's search intent. When the electronic device (100) determines that the search is successful because the target content has been selected, it may terminate the search session and obtain information on the search terms included in the search session and the search queries corresponding to the search terms.

[0088] For example, when a search term N is entered, a search result N is provided, and when content within the search result is selected by the user from the search result N, the electronic device (100) determines that the selected content is the target content and can obtain information about search terms included in the search session {search term A, search term B, search term C, ..., search term N-1, search term N} and search queries corresponding to each search term.

[0089] Among the search queries within the search session, previous search results A through N-1, where the target content was not selected, correspond to search failure cases, and search result N, where the target content was selected, corresponds to search success cases.

[0090] The electronic device (100) can analyze search results (hereinafter referred to as previous search results) (500) corresponding to previous search queries in which target content was not selected among search queries within a search session. For example, the electronic device (100) can analyze search result A, search result B, search result C, ..., and search result N-1, which are the previous search results (500).

[0091] For example, the cause of the search failure of each search result can be classified by identifying whether the target content is included in each search result. Specifically, the electronic device (100) can determine whether the search result is broad or narrow by identifying whether the target content is included in the search result but no content selection was made, or whether the target content is not included in the search result.

[0092] The electronic device (100) classifies the type of search query that causes the search result according to the cause of the search failure and can modify the knowledge graph based on the type of search query. For example, the electronic device (100) can add hidden connections within the knowledge graph so that the target content can be provided as a search result even when search queries corresponding to previous search results (500) occur. Hidden connections can serve to provide a new search path within the knowledge graph. When a search is optimized by adding hidden connections to the knowledge graph, the hidden connections of the knowledge graph can be used for subsequent searches. This is described with reference to FIG. 7.

[0093] FIG. 6 is a diagram illustrating the operation of an electronic device analyzing search results according to one embodiment of the present disclosure.

[0094] Referring to Fig. 6, examples of various search results based on a user's search input are illustrated.

[0095] For example, when a user inputs a search term A (610), the electronic device (100) can generate a search query corresponding to the search term A (610) and output a search result A (615). The search result A (615) may include target content (600), but may also include multiple contents in addition to the target content (600), so that the user may not be aware that the target content (600) is included in the search result A (615).

[0096] Additionally, for example, when a user inputs a search term B (620), the electronic device (100) may output a search result B (625) based on a search query corresponding to the search term B (625). The search result B (625) may not include target content (600).

[0097] Additionally, for example, when a user inputs a search term N (630), the electronic device (100) can output a search result N (635) based on a search query corresponding to the search term N (630). In this case, the search result N (635) includes target content (600) and other content other than the target content (600), and since the number of search results is appropriate, the user can recognize that the target content (600) corresponding to the search intent is included in the search results.

[0098] In other words, the user can sequentially input search terms A (610), B (620), etc., to find the target content (600), and only after inputting search term N (630) and reaching search result N (635) can the user discover and select the target content (600). In the examples described above, search result A (615) and search result B (625) correspond to search failure cases, and search result N (635) corresponds to search success cases.

[0099] In one embodiment, the electronic device (100) can analyze search failure cases. For example, the electronic device (100) can analyze search results corresponding to previous search queries. The electronic device (100) can identify whether the target content (600) is included in the search results corresponding to the previous search queries.

[0100] The electronic device (100) can classify search queries based on the results of analyzing whether the target content (600) is included in the search results. For example, the electronic device (100) can classify the search query as a broad search query if the target content is included in the search results, and classify the search query as a narrow search query if the target content (600) is not included in the search results.

[0101] Specifically, a search query A that occurs as search term A (610) is entered results in a search result A (615). Since search result A (615) is a search failure case and the search result includes target content (600), the electronic device (100) can classify search query A as a broad search query. Additionally, a search query B that occurs as search term B (620) is entered results in a search result B (625). Since search result B (625) is a search failure case and the search result does not include target content (600), the electronic device (100) can classify search query B as a narrow search query.

[0102] In one embodiment, the electronic device (100) may add hidden connections within a knowledge graph to optimize a search path based on the classification of a search query. The conditions under which the electronic device (100) adds hidden connections according to the classification of a search query may be predefined. The operation of the electronic device (100) determining the defined conditions for adding hidden connections will be further described in FIGS. 8, 9, and 11.

[0103] FIG. 7 is a diagram illustrating the operation of an electronic device according to one embodiment of the present disclosure to perform a search using a knowledge graph including hidden connection lines.

[0104] In one embodiment, a hidden connection line (700) added to the knowledge graph is set to a disabled state. The hidden connection line (700) of the knowledge graph can be activated when a preset condition is met. When the hidden connection line (700) is activated, the hidden connection line can perform the same function as other general connections of the knowledge graph.

[0105] At another point in time after a hidden connection line has been added to the knowledge graph, if a user executes a search function, the electronic device (100) may provide search results based on the user's search term (710) input. In this case, if the search term (710) corresponds to any one of the search terms included in the search session (720), the electronic device (100) may activate the hidden connection line (700) of the knowledge graph. For example, when a search is performed at a new point in time, if the search term (710) corresponds to at least one of the search terms within the search session (720), namely search term A, search term B, search term C, ..., search term N-1, search term N, the electronic device (100) may activate the hidden connection line (700) of the knowledge graph.

[0106] As described above in FIG. 5, among the search terms of the search session (720) of FIG. 7, search term A, search term B, search term C, ..., search term N-1 are search failure cases, and search term N is a search success case. For example, for search terms A, search term B, search term C, ..., search term N-1, which are search failure cases, the target content (730) may not be displayed in the search results, or too many search results may be displayed, making it difficult to recognize whether the target content (730) was searched.

[0107] When a hidden connection line (700) of the knowledge graph is activated, when any search term within the search session (720) is entered, the target content (730) can be provided as a search result through a new search path that includes the hidden connection line (700). In other words, when the user performs a search again later, the target content (730) is provided as a search result even for search terms within the search session (720) that were previously searched but failed, so that the user can select the target content (730) early, thereby increasing search efficiency.

[0108] FIG. 8 is a diagram illustrating the operation of adding a hidden connection line to a knowledge graph using an electronic device according to one embodiment of the present disclosure.

[0109] Referring to FIG. 8, search result A (810) is the search range when search term A is entered, and search result N (820) is the search range when search term N is entered.

[0110] When a search term is entered, the electronic device (100) can convert the search term to generate a search query and perform a search within the knowledge graph. The search range that is searched when a search term is entered may correspond to a set of nodes classified to represent a unit of user experience. The electronic device (100) may generate a set of nodes by analyzing interrelated nodes. For example, the electronic device (100) may designate nodes related to user actions and activities over a predetermined period as a set of nodes representing a unit of experience. For example, if search term A is "travel," a set of nodes related to the user's travel experience may correspond to search result A (810). And, if search term N is "domestic family trip," a set of nodes related to the user's experience of traveling domestically with family may correspond to search result N (820).

[0111] In the example of FIG. 8, search term A (e.g., "travel") corresponds to a search failure case where the target content (800) is included in the search results, but the search range is too vast, making it difficult to recognize the target content (800) within search result A (810). Subsequently, the user may enter additional search terms (e.g., "family trip," "family photo," "Seoul city travel," etc.) to find the desired target content (800), but the desired target content (800) may still not be provided as a search result because the search range is too vast or too narrow.

[0112] And, search term N (e.g., "domestic family trip") corresponds to a search success case in which the target content (800) is included in the search results and the search range is appropriate so that the target content (800) can be recognized within the search result N (820). In this case, the search query A converted from search term A can be classified as a broad search query by the electronic device (100).

[0113] The electronic device (100) can generate hidden connections (830) that connect sets of nodes within a knowledge graph to provide new search paths. For example, the electronic device (100) can generate hidden connections (830) that connect a set of nodes corresponding to search term A (illustrated as search result A (810)) and a set of nodes corresponding to search term N (illustrated as search result N (820)). The generated hidden connections are set to a deactivated state and can be activated when search term A is entered in a subsequent search.

[0114] FIG. 9 is a diagram illustrating the operation of adding a hidden connection line to a knowledge graph using an electronic device according to one embodiment of the present disclosure.

[0115] Referring to FIG. 9, search terms A, C, and D each illustrate examples where the search results include target content (900), but the search results are too broad. That is, search queries A, C, and D generated from search terms A, C, and D are classified into broad search queries by the electronic device (100).

[0116] The electronic device (100) can determine defined conditions for adding hidden connections based on the classification of the search query. For example, if the classification of the search query is a broad search query, the electronic device (100) can identify search paths corresponding to the broad search queries.

[0117] The electronic device (100) can identify a common starting point node where the starting points of the search overlap or an intersection point node where the search paths meet among the search paths of the wide search queries.

[0118] For example, when search term A is entered, the electronic device (100) searches the knowledge graph along the connecting lines from node A (910) to the target node (930). And when search term D is entered, the electronic device (100) also searches the knowledge graph along the connecting lines from node A (910) to the target node (930). In this case, node A (910) is the common starting point for the search path from search term A and the search path from search term D.

[0119] The electronic device (100) can identify a target node (930) corresponding to the target content (900) and create a hidden connection line connecting a set of nodes including a common starting node, node A (910), and a set of nodes including the target node (930). For example, the electronic device (100) can add a hidden connection line connecting the common starting node, node A (910), and the target node (930) to the knowledge graph.

[0120] Additionally, for example, when a search term C is entered, the electronic device (100) explores the knowledge graph along the connecting lines from the starting node corresponding to the search term C to the target node (930). In this case, node B (920) is the intersection point between the search path from search term A or search term D and the search path from search term C.

[0121] The electronic device (100) can identify a target node (930) corresponding to the target content (900) and generate a hidden connection line connecting a set of nodes including a node B (920), which is an intersection node, and a set of nodes including the target node (930). For example, the electronic device (100) can add a hidden connection line connecting the node B (920), which is an intersection node, and the target node (930) to a knowledge graph.

[0122] In one embodiment, in a search after a hidden connection line has been added to the knowledge graph, when at least one of search term A, search term C, or search term D is input into the electronic device (100), the electronic device (100) can activate a hidden connection line corresponding to the search term and provide search results using the knowledge graph. When a hidden connection line is created, the electronic device (100) can reduce the number of hops to search the knowledge graph. As a result, the target content (900) can be provided as a search result while the search range is reduced.

[0123] FIG. 10 is a diagram illustrating the operation of an electronic device according to one embodiment of the present disclosure providing search results using a knowledge graph including hidden connection lines.

[0124] Referring to FIG. 10, search result A (1010) is a search result when search term A is entered based on the original knowledge graph, and search result A' (1020) is a search result when search term A is entered based on the knowledge graph containing hidden connections.

[0125] Search query A generated from search term A is a broad search query, and when a search is performed based on the original knowledge graph, the search result A (1010) contains the target content (1000), but due to the numerous search results, it is difficult for the user to recognize the target content (1000). Therefore, the electronic device (100) can update the original knowledge graph by adding the aforementioned hidden connection line to the knowledge graph in order to provide the target content (1000) as a top item of the search results.

[0126] The electronic device (100) can provide search results based on the knowledge graph with hidden connections added when a wide search query is entered at a time after hidden connections have been added to the original knowledge graph.

[0127] The electronic device (100) may provide search results based on a knowledge graph with added hidden connections. For example, search result A' (1020) may include a first search result (1022) searched via a search path through hidden connections and a second search result (1024) searched via other search paths. In this case, the first search result (1022) may include target content (1000), and the first search result (1022) may be provided with higher priority than the second search result (1024).

[0128] In one embodiment, the electronic device (100) may obtain a first search result (1022) using a knowledge graph with added hidden connections and obtain a second search result (1024) using the original knowledge graph. In this case, the second search result (1024) may be the result of removing content that overlaps with the first search result (1022) from the result of performing a search based on the original knowledge graph.

[0129] In one embodiment, when the electronic device (100) uses a knowledge graph with added hidden connections, it can reduce the number of hops searched for a search. Accordingly, search result A' (1020) may include a first search result (1022) searched via a search path through hidden connections, and a second search result (1024) searched via a search path through connections other than hidden connections via the reduced number of hops. In this case, the second search result (1024) may contain fewer content than search result A (1010) because the search range is reduced due to the reduced number of hops.

[0130] FIG. 11 is a diagram illustrating the operation of adding a hidden connection line to a knowledge graph using an electronic device according to one embodiment of the present disclosure.

[0131] Referring to FIG. 11, search result B (1110) is the search range when search term B is entered, and search result N (1120) is the search range when search term N is entered.

[0132] In the example of FIG. 11, search term B corresponds to a search failure case where the target content (1100) is not included in the search results. Afterwards, the user may enter additional search terms to find the target content (1100), and the search range may still be too vast or too narrow, so the desired target content (1100) may not be provided as a search result.

[0133] And, search term N corresponds to a search success case in which the target content (1100) is included in the search result and the search range is appropriate so that the target content (1100) can be recognized within the search result N (1120). In this case, the search query B converted from search term B can be classified as a narrow search query by the electronic device (100).

[0134] The electronic device (100) can generate hidden connections (1130) that connect sets of nodes within a knowledge graph to provide new search paths. For example, the electronic device (100) can generate hidden connections (1130) that connect a set of nodes corresponding to search term B (illustrated as search result B (1110)) and a set of nodes corresponding to search term N (illustrated as search result N (1120)).

[0135] In one embodiment, when there are multiple sets of nodes corresponding to search term B, the electronic device (100) may select one of the multiple sets of candidate nodes. For example, the electronic device (100) may compare the set of nodes corresponding to search term N containing target content (1100) with the multiple sets of candidate nodes. Based on the result of comparing the attributes of the multiple sets of candidate nodes (e.g., similarity between entities, similarity of the user experience unit period included in the node set information, etc.) with the set of nodes containing target content (1100), the electronic device (100) may select one of the sets of candidate nodes and generate a hidden connection line connecting the two sets of nodes. The generated hidden connection line is set to a deactivated state and can be activated when search term B is entered in a subsequent search.

[0136] In one embodiment, the electronic device (100) can modify a narrow search query. For example, the electronic device (100) can determine a keyword (or sentence) by comparing a set of nodes corresponding to search term B with a set of nodes corresponding to search term N. The electronic device (100) can modify the search query so that the determined keyword is included in the narrow search query corresponding to search term B. Specifically, the electronic device (100) can modify the search query so that the node information of a third node (1140) in the knowledge graph is added as 'OR' to the search query B corresponding to search term B. Accordingly, even if a user enters search term B, the information of the third node (1140) is included in the search query generated from search term B, so the target content (1100) can be searched through a search path using the knowledge graph.

[0137] FIG. 12 is a diagram illustrating the operation of an electronic device according to one embodiment of the present disclosure to generate a knowledge graph with hidden connection lines added corresponding to a search session.

[0138] In one embodiment, adding hidden connections to the knowledge graph can be performed independently for each search session. Referring to FIG. 12, the electronic device (100) can store search sessions associated with the knowledge graph.

[0139] For example, the first search session (1210) may include N search terms such as search terms A1, A2, A3, ..., An. The electronic device (100) may perform a search result analysis for the first search session (1210) to generate a first knowledge graph (1215) with one or more hidden connections added, and may store information of the search session-knowledge graph by associating the first search session (1210) with the first knowledge graph (1215).

[0140] Additionally, for example, the electronic device (100) may perform an analysis of search results for a second search session (1220) to generate a second knowledge graph (1225) with one or more hidden connections added, and may store the second search session (1220) and the second knowledge graph (1225) in association. In the same way, the electronic device (100) may perform an analysis of search results for a third search session (1230) to generate a third knowledge graph (1235) with one or more hidden connections added, and may store the third search session (1230) and the third knowledge graph (1235) in association.

[0141] In one embodiment, the original knowledge graphs of the first knowledge graph (1215), the second knowledge graph (1225), and the third knowledge graph (1235) may be identical. The electronic device (100) may generate different hidden connections for each search session to generate a knowledge graph with hidden connections added corresponding to each search session. Since the electronic device (100) independently analyzes search results for search terms within each search session and adds hidden connections to the knowledge graph for each search session, results may appear in which different hidden connections are added to the same original knowledge graph. In one embodiment, the original knowledge graphs of the first knowledge graph (1215), the second knowledge graph (1225), and the third knowledge graph (1235) may be different. For example, if the subject of the search session is different, different knowledge graphs corresponding to each subject may be obtained as the original knowledge graph.

[0142] In one embodiment, the electronic device (100) can determine whether each search term entered in a search session is a search term within a single search session. For example, when the electronic device (100) receives search terms from a user, the user may enter a separate search term unrelated to the search intent in the middle of the search session. The electronic device (100) can analyze the association between the search terms within the search session and the target content based on the determination of the target content, and determine that search terms with an association below a threshold are not search terms within the search session. The electronic device (100) can store associated knowledge graph information for a search session consisting only of search terms determined to be search terms within a single search session. In this case, the associated knowledge graph may have one or more hidden connections added based on the analysis of search results of the search terms within the search session.

[0143] FIG. 13 is a diagram illustrating the operation of an electronic device according to one embodiment of the present disclosure to build a knowledge graph-based personalized database.

[0144] In one embodiment, the electronic device (100) can construct a knowledge graph-based personalized database by converting various types of unstructured data into the form of a knowledge graph. The electronic device (100) can acquire various types of data and convert them into triple-format data. The electronic device (100) can map the converted data to an ontology stored in semantic memory and reflect structured data that matches the ontology format into the knowledge graph. An ontology is a type of dictionary that defines terms conceptualizing data and the relationships between terms. An ontology can be expanded by adding external knowledge converted into triple-format data.

[0145] Data obtained from an electronic device (100) according to one embodiment of the present disclosure may be user data related to an application executed on the electronic device (100) or metadata regarding the time, place, weather, etc. of an event occurrence, stored on the electronic device (100). Data obtained from the electronic device (100) may include at least one of data input by a user on the electronic device (100), data detected by the electronic device (100), data received from an external source by the electronic device (100), and data processed by the electronic device (100).

[0146] The electronic device (100) can process unstructured data obtained from the electronic device (100) into structured data and store it in a knowledge graph. The electronic device (100) can prepare structured semantic information in advance as an ontology. The electronic device (100) can refer to semantic information of various types of ontology to process unstructured data obtained from the electronic device (100) into structured data and store it in a personalized database. For example, the electronic device (100) can convert various types of unstructured data into the form of a knowledge graph and store it in the knowledge graph.

[0147] Referring to FIG. 13, the process of an electronic device (100) collecting user data related to an application executed on the electronic device (100) and building a personalized database in the form of a knowledge graph is illustrated as an example.

[0148] In the data collector part of the electronic device (100), a user data collector (1330) can collect user data from a contact provider, a message provider, a media provider, CMH (content management hub) data, other data providers, etc.

[0149] Each provider may store raw data at the level provided by at least one application. For example, a message provider may store user data in the form of raw data received from a text message application or a chat application. The user data collector (1330) is not limited to the example shown in FIG. 13 and may collect user data from various types of applications, including third-party applications installed on the electronic device (100).

[0150] For example, a place data collector (1310) and a weather data collector (1320) can collect metadata regarding a place or weather from external information obtained from an electronic device (100). In FIG. 13, the place data collector (1310), the weather data collector (1320), and the user data collector (1330) are separated, but are not limited thereto, and the collectors listed as examples may be implemented in a form integrated into a single collector.

[0151] Each data collector (1310, 1320, 1330) collects data at the level of log data or raw data. For example, a user data collector (1330) can collect photo files themselves from a photo management application that manages photos taken through a camera. The user data collector (1330) can receive user data transmitted by each application according to a protocol predetermined by domain or by provider. When an event such as the addition, modification, or deletion of user data stored in each application occurs, the user data collector (1330) can receive the added user data, modified user data, or deleted user data from each application.

[0152] In the memory core part of the electronic device (100), the analyzer (1340) can generate a new form of user data using part or all of the data collected by each data collector (1310, 1320, 1330).

[0153] The analyzer (1340) can generate user data at the level of action information or activity information from the data collected by each of the data collectors (1310, 1320, 1330). The analyzer (1340) can generate user data at the level of action information or activity information by inferring and determining the user's act of taking a photo from a photo, the user's act of eating food from a receipt paid at a restaurant, and the user's act of staying from GPS-based location information of the user's location.

[0154] Action information is user data corresponding to a user's action (unit behavior), while activity information can be user data composed of a series of continuous action information. Activity information can have attribute data where the start and end times differ, whereas action information can have attribute data regarding the value of the time at which the user action occurred. For example, watching a movie can be considered activity information because the time the movie starts and the time it ends are different, whereas making a payment can be considered action information because the payment time is instantaneous. However, action information can also be treated as a type of activity information where the start and end times are the same.

[0155] The analyzer (1340) can generate user data at the knowledge graph level. The analyzer (1340) can generate a knowledge graph by obtaining action information or activity information based on data collected by each data collector (1310, 1320, 1330), connecting entities related to the action information or activity information as nodes, and storing attribute information of each node. For example, the knowledge graph may be composed of a lower layer containing user data at the raw data level, a middle layer containing user data at the action information or activity information level, and an upper layer corresponding to the title or topic level of the knowledge graph encompassing the action information or activity information.

[0156] Multiple nodes corresponding to activity information included in the middle layer can be connected to the top layer of the knowledge graph. Each node corresponding to activity information can be connected as attribute data of each activity information by multiple nodes corresponding to user data at the raw data level included in the lower layer. Nodes corresponding to activity information in the knowledge graph can be connected to each other, and nodes corresponding to user data can be connected to nodes corresponding to other user data.

[0157] Nodes corresponding to similar user data contained in one or more knowledge graphs can be clustered and managed as a single group.

[0158] For example, photos containing the faces of look-alike users are managed as a single group, so that when attribute information is entered for any one of the photos, all photos within the same group can share the entered attribute information. If any one photo within the same group is deleted, all photos within the same group can be processed to be deleted from one or more knowledge graphs.

[0159] As another example, a knowledge graph can represent user data generated within a specified period as a root node, with the specified period information serving as the root node and each node connected to the root node. The root node is traversable through the specified period information. The root node can be connected to each knowledge graph, with each sub-period information constituting the specified period information serving as a sub-root node.

[0160] The encoder (1350) converts user data generated or processed by the analyzer (1340) into triple-format data and maps it to a predetermined ontology, infers structured data that fits the ontology format, and can build a personal knowledge graph in storage (1360) or reflect it in an existing knowledge graph. Various types of ontology, such as content ontology, activity information ontology, environment ontology, and relationship ontology, may be prepared in advance in the electronic device (100).

[0161] The encoder (1350) can integrate and use various ontologies prepared in the electronic device (100). For example, the encoder (1350) can convert user data related to card payment information made on January 1 and photos taken on January 1 into a triple format, and use an activity information ontology to identify the converted user data as activity information for purchases made on January 1 and activity information for taking photos, respectively, and generate or update a knowledge graph based on activity information in a structured form.

[0162] Storage (1360) can store a knowledge graph. The electronic device (100) can manage the knowledge graph by building a knowledge graph-based personalized database in the storage (1360). The electronic device (100) can record and store data related to events generated by users in the form of a knowledge graph over a certain period. The electronic device (100) can store and manage a knowledge graph in which objects related to events are nodes, in order to record and manage events generated in the electronic device (100) using various types of user data.

[0163] FIG. 14 is a drawing for exemplarily illustrating the use of a knowledge graph generated in an electronic device according to one embodiment of the present disclosure.

[0164] Referring to FIG. 14, an electronic device (100) according to one embodiment of the present disclosure can collect user data obtained from the electronic device (100) and generate a knowledge graph with the theme of "restaurant tour." Based on the collected user data, the electronic device (100) can generate a node of activity information called "staying," a node of activity information called "eating," and a node of activity information called "taking pictures," and can generate a knowledge graph called "pizza restaurant" by connecting each node of activity information. The electronic device (100) can store and manage the knowledge graph generated in this way in a knowledge graph-based personalized database.

[0165] Since each node of the activity information constituting the knowledge graph stores attribute data regarding the start time and end time, the electronic device (100) can use the knowledge graph to identify whether there is an overlapping time between each activity information. If there is overlapping activity information at the same time, the electronic device (100) can utilize attribute data regarding one activity information as attribute data regarding another activity information. The electronic device (100) can infer new information by using a predetermined inference model to integrate and use user data at the raw data level of multiple activity information from activity information where the performed times overlap.

[0166] For example, referring to Fig. 14, the activity information 'Staying' may include information that the user was at a place called Italian Table in Gangnam-gu, Seoul from 2:50 PM to 3:50 PM on May 18. The activity information 'Eating' may include information that the user had a meal with friend Daniel from 3:00 PM to 3:50 PM on May 18. The activity information 'Taking Photos' may include information that the user took several photos related to pizza from 3:20 PM to 3:25 PM on May 18.

[0167] The electronic device (100) can infer from the knowledge graph that the user took a picture of the pizza while having a meal with his friend Daniel at a place called Italian Table in Gangnam-gu, Seoul, from 3:20 PM to 3:25 PM on May 18, the time the pizza was photographed.

[0168] Therefore, when a user searches for a place where pizza was eaten or a person who ate pizza with on an electronic device (100), by utilizing a knowledge graph called 'Pizza Restaurant', it can be identified that the place where pizza was eaten is 'Italian Table in Gangnam-gu, Seoul' and the person who ate pizza with is 'Friend Daniel'.

[0169] FIG. 15 is a drawing for illustrating an example of a knowledge graph generated in an electronic device according to one embodiment of the present disclosure.

[0170] An electronic device (100) according to one embodiment can collect user data related to each application and classify it according to a predetermined criterion. For example, the electronic device (100) can classify user data by clustering user data collected at similar times or in similar locations. The electronic device (100) can identify user action information or activity information from user data related to various applications using an action information ontology or an activity information ontology. The electronic device (100) can obtain a knowledge graph in which user action information or activity information is represented as each node.

[0171] The electronic device (100) can generate a knowledge graph by connecting nodes of activity information for a specific period. Referring to the knowledge graph illustrated in FIG. 15, it can be seen that among the user data related to each application, nodes of activity information generated based on user data with a generation time of May 18 are connected.

[0172] For example, the knowledge graph may include a node of activity information named 'Watch Video', which is generated based on user data related to the title, artist information, playback start time, and playback end time of a video played through a content playback application.

[0173] For example, the knowledge graph may include a node of activity information called 'meeting', generated based on user data related to the title, start time, and end time of a schedule recorded through a schedule management application.

[0174] For example, the knowledge graph may include nodes of activity information called 'staying,' generated based on user data related to location addresses and points of interest recorded through a GPS-based application.

[0175] For example, the knowledge graph may include a node of activity information called 'calling', generated based on user data related to the person called through the phone application, the call start time, and the call end time.

[0176] For example, the knowledge graph may include a node of activity information called 'taking pictures,' which is generated based on user data related to photos taken through a photo management application, the subject, the start time, and the end time of the shoot, as well as user data related to counterpart information stored in a contact application.

[0177] For example, the knowledge graph may include a node of activity information called 'purchasing', generated based on user data related to payment information (payment recipient, payment card, payment amount) and payment time through a messaging application.

[0178] Referring to the examples above, it can be seen that the activity information of each node in the knowledge graph is related to the activity information that took place on May 18.

[0179] The electronic device (100) can provide a service and personalized experience tailored to the individual user by utilizing a knowledge graph in the service. As illustrated in FIG. 15, if a knowledge graph based on mutually correlated user data is utilized in the service, the electronic device (100) can provide a service of improved quality with noise removed or provide a new type of service based on mutually correlated user data. Below, a method for generating a knowledge graph based on mutually correlated user data is described.

[0180] FIG. 16 is a flowchart illustrating the operation of an electronic device generating a knowledge graph according to one embodiment of the present disclosure.

[0181] Referring to FIG. 16, the electronic device (100) can obtain a first group of clustered data from user data associated with a first application (1610). The first application is an application installed on the electronic device (100) and may include, but is not limited to, a schedule management application, a chat application, a photo management application, a GPS-based application (e.g., a map application), an SNS application, a health information management application, a message application, a phone application, a weather forecast application, a content playback application, etc.

[0182] The first data group may include one or more user data. For example, if the user data is a chat message, the electronic device (100) may cluster the chat messages based on information regarding the time of creation of the chat message or information regarding the chat partner. The electronic device (100) may acquire chat messages for a predetermined period of time as the first data group based on the time of creation of each chat message. The electronic device (100) may acquire chat messages with a specific partner as the first data group based on information regarding the chat partner.

[0183] For example, if user data is a photograph, the electronic device (100) can cluster the photograph based on at least one of information regarding the time of creation (time taken) of the photograph, information regarding the place where the photograph was taken, and information regarding the object that was taken. The electronic device (100) can acquire photographs taken consecutively as a first data group based on the time when the photograph was taken. The electronic device (100) can acquire photographs taken at the same place as a first data group. The electronic device (100) can acquire photographs containing the same object as a first data group based on the object that was taken. The electronic device (100) can also acquire a single photograph as a first data group.

[0184] FIG. 17 is a diagram illustrating a first data group and a second data group clustered from user data related to a plurality of applications in an electronic device according to one embodiment of the present disclosure.

[0185] Referring to FIG. 17, user data generated in a schedule management application, a chat application, a photo management application, a GPS-based application, and a social media application are shown. For example, in a schedule-related application, schedule information can be generated when a user records a schedule. In a chat application, chat messages can be generated when a user chats with another party. In a photo management application, photos or videos generated when a user takes photos or videos can be stored. In a GPS-based application, location information can be generated at predetermined time intervals. In a social media application, a social media feed can be generated when a user updates content on social media.

[0186] As illustrated in FIG. 17, in the case of chat messages related to a chat application, a user may continuously exchange chat messages with another party, or intermittently or one-time messages may be transmitted to or received by the electronic device (100). In the case of chat messages exchanged continuously or chat messages with a specific counterpart, it is highly likely that they are user data with mutual correlation. The electronic device (100) can obtain a first chat message data group and a second chat message data group by clustering chat messages that are continuously generated or chat messages exchanged with a specific counterpart over a certain period among the chat messages.

[0187] In the case of photos or videos related to a photo management application, a one-time photo or video may be taken at a specific moment or place, or it may be taken continuously or both photos and videos may be taken. Since a one-time shot may also be meaningful for photos and videos, it is necessary to process even a single photo as a data group. The electronic device (100) can obtain a first image data group and a second image data group by clustering one or more photos or videos.

[0188] In the case of schedule information related to a schedule management application, location information related to a GPS-based application, information regarding a social media feed related to a social media application, etc., the electronic device (100) can obtain a data group by clustering one or more user data.

[0189] The electronic device (100) can acquire any data group among clustered data groups as a first data group. For example, the electronic device (100) can acquire any data group clustered from user data related to any application as a first data group.

[0190] Referring again to FIG. 16, the electronic device (100) can determine a generation period corresponding to the first data group (hereinafter, the first generation period) based on the generation time of user data included in the first data group. (1620)

[0191] For example, the electronic device (100) can determine the period from the time of creation of the first user data among the user data belonging to the first data group to the time of creation of the last user data as the first creation period. Referring again to FIG. 17, the electronic device (100) can obtain a first chat message data group clustered from user data related to a chat application as the first data group. The electronic device (100) can determine the period from the time of creation of the first chat message among the chat messages belonging to the first chat message data group to the time of creation of the last chat message as the first creation period (T1_C).

[0192] For example, the electronic device (100) may identify a period corresponding to a predetermined time before and after the time of creation of user data belonging to the first data group as the first creation period. Referring again to FIG. 17, the electronic device (100) may acquire a second image data group clustered from user data related to a photo management application as the first data group. The electronic device (100) may determine a period corresponding to a predetermined time before and after the time of creation of photos belonging to the second image data group as the first creation period (T1_P).

[0193] Referring again to FIG. 16, the electronic device (100) can obtain at least one second data group corresponding to a first generation period, clustered from user data related to at least one second application different from the first application. (1630)

[0194] Referring again to FIG. 17, if the electronic device (100) acquires a first chat message data group as the first data group, it may acquire a first schedule data group, a first image data group, a first location information data group, and an SNS feed data group generated during a first generation period (T1_C) corresponding to the first chat message data group as the second data group. The generation period of each of the first schedule data group, the first image data group, the first location information data group, and the SNS feed data group generated during the first generation period (T1_C) corresponding to the first chat message data group does not exceed the range of the first generation period (T1_C) corresponding to the first chat message data group.

[0195] If the electronic device (100) acquires a second image data group as a first data group, it may acquire a second schedule data group, a second chat message data group, and a second location information data group generated during a first generation period (T1_P) corresponding to the second image data group as a second data group. The generation period of each of the second schedule data group, the second chat message data group, and the second location information data group generated during the first generation period (T1_P) corresponding to the second image data group does not exceed the range of the first generation period (T1_C) corresponding to the second image data group.

[0196] If the second generation period corresponding to the second data group falls outside the range of the first generation period corresponding to the first data group, the electronic device (100) may extend the second generation period corresponding to the second data group by extending the range of the first generation period. In this regard, this is described with reference to FIGS. 18 and 19.

[0197] FIG. 18 is a flowchart illustrating the process of expanding and acquiring a second data group in an electronic device according to one embodiment of the present disclosure.

[0198] Referring to FIG. 18, the electronic device (100) can extract user data corresponding to a first generation period corresponding to a first data group among user data related to at least one second application. (1810) If the user data related to the second application is continuous from before the start time and after the end time of the first generation period corresponding to the first data group, the electronic device (100) can extract continuous user data that extends beyond the first generation period corresponding to the first data group.

[0199] The electronic device (100) can obtain at least one second data group by clustering extracted user data related to at least one second application based on information regarding the time of creation. (1820) The at least one second data group may include at least one data group related to the second application and at least one data group related to the third application.

[0200] The electronic device (100) can determine whether a second generation period corresponding to at least one second data group exceeds the first generation period. (1830) The electronic device (100) can determine whether the period from the generation time of the first user data among the user data belonging to at least one second data group to the generation time of the latest user data exceeds the range of the first generation period corresponding to the first data group. The electronic device (100) can determine whether the range of the generation period of at least one data group related to the second application and the range of the generation period of at least one data group related to the third application exceed the range of the first generation period corresponding to the first data group related to the first application.

[0201] The electronic device (100) may extend the first generation period based on the second generation period if the second generation period corresponding to at least one second data group exceeds the first generation period. (1840) The electronic device (100) may repeat the process of acquiring at least one second data group for the extended first generation period. The electronic device (100) may terminate the process of acquiring at least one second data group if the second generation period corresponding to at least one second data group does not exceed the first generation period.

[0202] FIG. 19 is a drawing for illustrating the expansion of a second data group in an electronic device according to one embodiment of the present disclosure.

[0203] Referring to FIG. 19, the electronic device (100) can obtain a clustered image data group (first data group) from user data associated with a photo management application (first application). The electronic device (100) can identify a first generation period (T1) corresponding to the image data group. The electronic device (100) can obtain a schedule data group corresponding to the first generation period (T1), clustered from schedule information associated with a schedule management application. The electronic device (100) can obtain a first chat message data group and a second chat message data group corresponding to the first generation period, clustered from chat messages associated with a chat application. The electronic device (100) can obtain a first location information data group corresponding to the first generation period (T1), clustered from location information associated with a GPS-based application. Since there is no information regarding SNS feeds related to SNS applications corresponding to the first generation period (T1), the electronic device (100) cannot obtain an SNS feed data group corresponding to the first generation period (T1) from information regarding SNS feeds related to SNS applications. Accordingly, the electronic device (100) can obtain a schedule data group, a first chat message data group, a second chat message data group, and a first location information data group corresponding to the first generation period (T1) as a second data group.

[0204] At this time, it can be seen that the generation period corresponding to the first chat message data group and the generation period corresponding to the second chat message data group exceed the first generation period (T1) corresponding to the video data group. The electronic device (100) can extend the first generation period (T1) to a new first generation period (T2) based on the generation period corresponding to the first chat message data group and the generation period corresponding to the second chat message data group. The electronic device (100) can acquire at least one second data group corresponding to the extended first generation period (T2).

[0205] The electronic device (100) can obtain a schedule data group corresponding to an extended first generation period (T2) clustered from schedule information related to a schedule management application. The electronic device (100) can obtain a first chat message data group and a second chat message data group corresponding to an extended first generation period (T2) clustered from chat messages related to a chat application. The electronic device (100) can obtain a first location information data group and a second location data group corresponding to an extended first generation period (T2) clustered from location information related to a GPS-based application. The electronic device (100) can obtain a social media feed data group corresponding to an extended first generation period (T2) clustered from information regarding social media feeds related to a social media application. Consequently, as the existing first generation period (T1) changes to an extended first generation period (T2), the electronic device (100) can further obtain a second location information data group and a social media feed data group as a second data group.

[0206] Referring again to FIG. 16, the electronic device (100) can obtain a knowledge graph based on mutually correlated user data from a first data group and at least one second data group. (1640)

[0207] Interrelatedness refers to the formation or existence of a relationship between user data. For example, if 'User Data A' has a sequential or causal relationship with 'User Data B' based on a specific point in time, 'User Data A' and 'User Data B' may be recognized as having interrelatedness. If 'User Data C' and 'User Data D' possess common attributes, such as relating to the same subject or being generated in the same location, 'User Data C' and 'User Data D' may be recognized as having interrelatedness.

[0208] Knowledge graphs based on user data can be used to represent user characteristics, tendencies, preferences, or tastes, or to explain specific events related to the user. Knowledge graphs based on user data with recognized interrelationships can serve as a form of knowledge graph optimized for providing knowledge graph-based personalized services.

[0209] FIG. 20 is a flowchart illustrating the process of obtaining a knowledge graph based on interrelated user data in an electronic device according to one embodiment of the present disclosure.

[0210] Referring to FIG. 20, the electronic device (100) can generate a knowledge graph based on a first data group and at least one second data group. (2010) The electronic device (100) can generate a knowledge graph by obtaining action information or activity information based on all user data included in the first data group and at least one second data group, connecting the action information or activity information to each other as nodes, and storing attribute information of each node.

[0211] The electronic device (100) can determine the mutual association of each node in a knowledge graph (2020). The electronic device (100) can calculate the mutual association of each node based on at least one of the following: the presence or absence of a node connected to a first node based on user data of a first data group (first determination criterion), the number of nodes connected to the first node (second determination criterion), the presence or absence of a node connected to a second node connected to the first node (third determination criterion), and the number of nodes connected to the second node (fourth determination criterion).

[0212] The electronic device (100) can obtain the mutual correlation of each node according to the output value of a predetermined function or a predetermined learning model that takes at least one of the values ​​quantified according to each judgment criterion as input. The predetermined function may be a function that outputs a value through a predetermined operation among the values ​​quantified according to the judgment criterion, or selects a value according to a predetermined rule. Alternatively, the predetermined function may be a function that calculates an average such as an arithmetic mean, harmonic mean, geometric mean, or weighted mean of the values ​​quantified according to the judgment criterion, or outputs a minimum or maximum value. Alternatively, the predetermined function may be a function that takes at least one of the values ​​quantified according to the judgment criterion as input and outputs a predetermined operation result. The predetermined learning model may be a mutual correlation determination model in the form of a deep learning model or a machine learning model that takes at least one of the values ​​quantified according to the judgment criterion as input. However, if the electronic device (100) fails to determine the values ​​quantified according to all judgment criteria, the electronic device (100) may determine a predetermined value as the value representing the mutual correlation.

[0213] The electronic device (100) can remove nodes that are not mutually related from the knowledge graph (2030). The electronic device (100) can remove nodes and nodes at a level lower than a predetermined standard that have a value indicating mutual relatedness from the knowledge graph.

[0214] The electronic device (100) can determine whether there is a singularity in the knowledge graph (2040). The electronic device (100) can generate sentences to be input into a large language model or a machine learning-based model that determines singularity from user data of each node reflected in the knowledge graph. The electronic device (100) can input the generated sentences into a large language model or a machine learning-based model that determines singularity to determine whether there is a singularity in the knowledge graph and to determine the topic or title of the knowledge graph.

[0215] The electronic device (100) may store a knowledge graph based on interrelated user data if there is a specificity in the knowledge graph (2050). Even if it is a knowledge graph based on interrelated user data, if it is determined that there is no specificity in the knowledge graph, the electronic device (100) may not store the knowledge graph.

[0216] The present disclosure relates to a method for providing content search to a user based on a knowledge graph and an electronic device for providing such a method. The technical problems to be solved by the present disclosure are not limited to those mentioned above, and other technical problems not mentioned will be clearly understood by those skilled in the art from the description in this specification.

[0217] According to one aspect of the present disclosure, a method may be provided in which an electronic device provides content retrieval.

[0218] The above method may include the step of providing knowledge graph-based search results during a search session based on search term input.

[0219] The above search session may include sequential search queries, and a corresponding search result may be provided whenever each search query occurs.

[0220] The above method may include the step of terminating the search session and identifying the selected content as target content when content is selected from the search results.

[0221] The above method may include a step of analyzing search results corresponding to previous search queries in which the target content was not selected among the search queries within the search session.

[0222] The above method may include the step of optimizing the knowledge graph-based search by adding hidden connections within the knowledge graph based on the analysis results to provide the target content as a main item in the search results corresponding to the previous search queries.

[0223] The knowledge graph above may include nodes representing entities related to the user of the electronic device and connecting lines representing the relationships between the entities.

[0224] The knowledge graph above may be constructed by defining the entities and the relationships between the entities based on data collected from the electronic device.

[0225] The knowledge graph above may include node set information in which nodes are classified to represent the user's experience unit.

[0226] The above hidden connections may be connected between sets of nodes within the knowledge graph to provide a new search path.

[0227] The above hidden connection line may be set to a disabled state.

[0228] The above method may include the step of activating the hidden connection line based on inputting a search term included in the search session.

[0229] The step of analyzing the above search results may include a step of identifying whether the target content is included in the search results corresponding to the above previous search queries.

[0230] The step of analyzing the above search results may include classifying the above previous search queries into broad search queries in which the target content is included in the search results or narrow search queries in which the target content is not included in the search results.

[0231] The step of optimizing the knowledge graph-based search may include determining a defined condition for adding the hidden connection line based on the classification of the search query.

[0232] The step of optimizing the knowledge graph-based search may include the step of identifying search paths corresponding to the broad search queries within the knowledge graph.

[0233] The step of optimizing the knowledge graph-based search above may include the step of identifying the intersection nodes of the search paths.

[0234] The step of optimizing the knowledge graph-based search may include the step of identifying a target node corresponding to the target content within the knowledge graph.

[0235] The step of optimizing the knowledge graph-based search may include the step of generating a hidden connection line that connects a set of nodes containing the intersection node and a set of target nodes containing the target node.

[0236] The step of optimizing the knowledge graph-based search may include reducing the number of hops to search the knowledge graph when a hidden connection line is created connecting a set of nodes containing an intersection node and a set of target nodes containing a target node.

[0237] The above method may include the step of providing search results based on a knowledge graph with added hidden connections based on inputting the above-described extensive search query.

[0238] The above search results may include a first search result searched via a search path through the above hidden connection line and a second search result searched via other search paths.

[0239] The above search result may be such that the first search result is provided with higher priority than the second search result.

[0240] The step of optimizing the knowledge graph-based search may include the step of identifying a set of nodes corresponding to the narrow search query within the knowledge graph.

[0241] The step of optimizing the knowledge graph-based search may include the step of identifying a target node corresponding to the target content within the knowledge graph.

[0242] The step of optimizing the knowledge graph-based search above may include the step of generating a hidden connection line connecting a set of nodes corresponding to the narrow search query and a set of target nodes including the target node.

[0243] The above method may include a step of determining whether each search term is a search term within a single search session for search terms entered in the search session.

[0244] The above method may include the step of storing a search session containing search terms determined as search terms within the single search session in association with a knowledge graph to which the hidden connection line has been added.

[0245] The above method may generate a knowledge graph with added hidden connections corresponding to each search session by generating different hidden connections for each search session.

[0246] According to one aspect of the present disclosure, an electronic device that provides content retrieval may be provided.

[0247] The electronic device may include at least one processor; a memory for storing instructions; and a display.

[0248] By executing the above instructions by the at least one processor, the electronic device can provide knowledge graph-based search results during a search session based on search term input.

[0249] The above search session may include sequential search queries, and a corresponding search result may be provided whenever each search query occurs.

[0250] By executing the above instructions by the at least one processor, the electronic device can terminate the search session and identify the selected content as target content when content is selected from the search results.

[0251] By executing the above instructions by the at least one processor, the electronic device can analyze search results corresponding to previous search queries in which the target content was not selected among the search queries within the search session.

[0252] By executing the above instructions by the at least one processor, the electronic device can optimize the knowledge graph-based search by adding hidden connections within the knowledge graph based on the analysis results to provide the target content as a main item in the search results corresponding to the previous search queries.

[0253] The knowledge graph above may include nodes representing entities related to the user of the electronic device and connecting lines representing the relationships between the entities.

[0254] The knowledge graph above may be constructed by defining the entities and the relationships between the entities based on data collected from the electronic device.

[0255] The knowledge graph above may include node set information in which nodes are classified to represent the user's experience unit.

[0256] The above hidden connections may be connected between sets of nodes within the knowledge graph to provide a new search path.

[0257] The above hidden connection line may be set to a disabled state.

[0258] By executing the above instructions by the at least one processor, the electronic device can activate the hidden connection line based on inputting a search term included in the search session.

[0259] By executing the above instructions by the at least one processor, the electronic device can identify whether the target content is included in the search results corresponding to the previous search queries.

[0260] The aforementioned previous search queries can be classified into broad search queries in which the target content is included in the search results or narrow search queries in which the target content is not included in the search results.

[0261] Based on the classification of the above search query, a defined condition for adding the above hidden connection line can be determined.

[0262] By executing the above instructions by the at least one processor, the electronic device can identify search paths corresponding to the extensive search queries within the knowledge graph.

[0263] By executing the above instructions by the at least one processor, the electronic device can identify the intersection node of the search paths.

[0264] By executing the above instructions by the at least one processor, the electronic device can identify a target node corresponding to the target content within the knowledge graph.

[0265] By executing the above instructions by the at least one processor, the electronic device can generate a hidden connection line connecting a set of nodes including the intersection node and a set of target nodes including the target node.

[0266] By executing the above instructions by the at least one processor, the electronic device can reduce the number of hops traversing the knowledge graph when a hidden connection line is created connecting a set of nodes including an intersection node and a set of target nodes including a target node.

[0267] By executing the above instructions by the at least one processor, the electronic device can provide search results based on a knowledge graph with added hidden connections based on the input of the broad search query.

[0268] The above search results may include a first search result searched via a search path through the above hidden connection line and a second search result searched via other search paths.

[0269] The above search result may be such that the first search result is provided with higher priority than the second search result.

[0270] By executing the above instructions by the at least one processor, the electronic device can identify a set of nodes corresponding to the narrow search query within the knowledge graph.

[0271] By executing the above instructions by the at least one processor, the electronic device can identify a target node corresponding to the target content within the knowledge graph.

[0272] By executing the above instructions by the at least one processor, the electronic device can generate a hidden connection line connecting a set of nodes corresponding to the narrow search query and a set of target nodes including the target node.

[0273] By executing the above instructions by the at least one processor, the electronic device can determine whether each search term is a search term within a single search session for search terms entered in the search session.

[0274] By executing the above instructions by the at least one processor, the electronic device can store a search session including search terms determined as search terms within the single search session in association with a knowledge graph to which the hidden connection line has been added.

[0275] By executing the above instructions by the at least one processor, the electronic device can generate a different hidden connection line for each search session and generate a knowledge graph with a hidden connection line added corresponding to each search session.

[0276] Meanwhile, embodiments of the present disclosure may also be implemented in the form of a recording medium containing computer-executable instructions, such as program modules executed by a computer. A computer-readable medium may be any available medium accessible by a computer and includes both volatile and non-volatile media, and both removable and non-removable media. Additionally, a computer-readable medium may include computer storage media and communication media. Computer storage media include both volatile and non-volatile, removable and non-removable media implemented by any method or technique for storing information, such as computer-readable instructions, data structures, program modules, or other data. Communication media may typically include other data of modulated data signals, such as computer-readable instructions, data structures, or program modules.

[0277] Additionally, computer-readable storage media may be provided in the form of non-transitory storage media. Here, 'non-transitory storage media' simply means that it is a tangible device and does not contain a signal (e.g., electromagnetic waves), and this term does not distinguish between cases where data is stored semi-permanently and cases where it is stored temporarily. For example, 'non-transitory storage media' may include a buffer in which data is stored temporarily.

[0278] According to one embodiment, the method according to the various embodiments disclosed herein may be provided by being included in a computer program product. The computer program product may be traded between a seller and a buyer as a product. The computer program product may be distributed in the form of a device-readable storage medium (e.g., compact disc read-only memory (CD-ROM)), or distributed online (e.g., download or upload) through an application store or directly between two user devices (e.g., smartphones). In the case of online distribution, at least a portion of the computer program product (e.g., a downloadable app) may be temporarily stored or temporarily created on a device-readable storage medium, such as the memory of a manufacturer's server, an application store's server, or a relay server.

[0279] The foregoing description of the present disclosure is for illustrative purposes only, and those skilled in the art will understand that other specific forms can be easily modified without altering the technical spirit or essential features of the present disclosure. Therefore, the embodiments described above should be understood as illustrative in all respects and not restrictive. For example, each component described as a single unit may be implemented in a distributed manner, and components described as distributed may likewise be implemented in a combined form.

[0280] The scope of the present disclosure is defined by the claims set forth below rather than by the detailed description above, and all modifications or variations derived from the meaning and scope of the claims and equivalent concepts thereof should be interpreted as being included within the scope of the present disclosure.

Claims

1. A method for an electronic device to provide content retrieval, Step (310), wherein a knowledge graph-based search result is provided during a search session based on search term input, wherein the search session includes sequential search queries and a corresponding search result is provided whenever each search query occurs; Step (320) of terminating the search session and identifying the selected content as target content when content is selected from the search results above; Among the search queries within the above search session, a step (330) of analyzing search results corresponding to previous search queries in which the target content was not selected; and A method comprising the step (340) of optimizing the knowledge graph-based search by adding hidden connections within the knowledge graph based on the analysis results to provide the target content as a main item in the search results corresponding to the previous search queries.

2. In Paragraph 1, A method wherein the knowledge graph comprises nodes representing entities related to a user of the electronic device and connecting lines representing the relationships between the entities, is constructed by defining the entities and the relationships between the entities based on data collected from the electronic device, and includes node set information in which the nodes are classified to represent the user's experience unit.

3. In Paragraph 2, A method in which the above hidden connections are connected between sets of nodes within the knowledge graph to provide a new search path.

4. In Paragraph 3, The above hidden connection line is set to a disabled state, and The above method is, A method further comprising the step of activating the hidden connection line based on inputting a search term included in the above search session.

5. In any one of paragraphs 2 through 4, The step of analyzing the above search results is, A step of identifying whether the target content is included in the search results corresponding to the aforementioned previous search queries; and The method includes the step of classifying the aforementioned prior search queries into broad search queries in which the target content is included in the search results or narrow search queries in which the target content is not included in the search results, and The step of optimizing the above knowledge graph-based search is, A method comprising the step of determining a defined condition for adding the hidden connection line based on the classification of the above search query.

6. In Paragraph 5, The step of optimizing the above knowledge graph-based search is, A step of identifying search paths corresponding to the broad search queries within the knowledge graph; A step of identifying the intersection node of the above search paths; A step of identifying a target node corresponding to the target content within the knowledge graph; and A method comprising the step of generating a hidden connection line connecting a set of nodes including the intersection node and a set of target nodes including the target node.

7. In Paragraph 5, The step of optimizing the above knowledge graph-based search is, A step of identifying a set of nodes corresponding to the narrow search query within the knowledge graph; A step of identifying a target node corresponding to the target content within the knowledge graph; and A method comprising the step of generating a hidden connection line connecting a set of nodes corresponding to the above narrow search query and a set of target nodes including the target node.

8. In an electronic device (100) that provides content search, At least one processor (110); Memory (120) for storing instructions; and Includes a display (130), By executing the above instructions by the at least one processor (110), the electronic device (100) Knowledge graph-based search results are provided during a search session based on search term input, wherein the search session includes sequential search queries, and a corresponding search result is provided whenever each search query occurs. When content is selected from the above search results, the above search session is terminated and the selected content is identified as target content, and Among the search queries within the above search session, analyze the search results corresponding to previous search queries in which the target content was not selected, and An electronic device (100) that optimizes the knowledge graph-based search by adding hidden connection lines within the knowledge graph based on the analysis results to provide the target content as a main item in the search results corresponding to the previous search queries.

9. In Paragraph 8, The electronic device, wherein the knowledge graph comprises nodes representing entities related to the user of the electronic device and connecting lines representing the relationships between the entities, is constructed by defining the entities and the relationships between the entities based on data collected from the electronic device, and includes node set information classifying the nodes to represent the user's experience unit.

10. In Paragraph 9, An electronic device in which the above hidden connection lines connect between sets of nodes within the knowledge graph to provide a new search path.

11. In Paragraph 10, The above hidden connection line is set to a disabled state, and By executing the above instructions by the at least one processor, the electronic device, An electronic device that activates the hidden connection line based on inputting a search term included in the above search session.

12. In any one of paragraphs 8 through 10, By executing the above instructions by the at least one processor, the electronic device, Identify whether the target content is included in the search results corresponding to the aforementioned previous search queries, and The aforementioned previous search queries are classified into broad search queries in which the target content is included in the search results or narrow search queries in which the target content is not included in the search results, and An electronic device that determines defined conditions for adding the hidden connection line based on the classification of the above search query.

13. In Paragraph 12, By executing the above instructions by the at least one processor, the electronic device, Identify search paths corresponding to the above-mentioned extensive search queries within the above-mentioned knowledge graph, and Identify the intersection node of the above search paths, and Identify the target node corresponding to the target content within the knowledge graph above, and An electronic device that generates a hidden connection line connecting a set of nodes including the above-mentioned intersection node and a set of target nodes including the above-mentioned target node.

14. In Paragraph 12, By executing the above instructions by the at least one processor, the electronic device, Identifying a set of nodes corresponding to the narrow search query within the knowledge graph, and Identify the target node corresponding to the target content within the knowledge graph above, and An electronic device that generates a hidden connection line connecting a set of nodes corresponding to the above narrow search query and a set of target nodes including the target node.

15. A computer-readable recording medium having a program for executing the method of any one of paragraphs 1 through 7 on a computer.

Images