Method and apparatus for trend-based content generation

The content generation device addresses the limitations of conventional methods by normalizing and evaluating trend signals for lifecycle and risk, generating content guidelines, and verifying differentiation, resulting in safe and aligned content publication.

KR103043693B1Active Publication Date: 2026-11-06EN HANCE
View PDF 3 Cites 0 Cited by

Patent Information

Application Number
KR1020260132511
Authority / Receiving Office
KR · KR
Patent Type
Patents
Current Assignee / Owner
Filing Date
2026-07-20
Publication Date
2026-11-06
Estimated Expiration
2046-07-20

AI Technical Summary

Technical Problem

Conventional content generation methods fail to consider the meaning, context, risk, lifecycle, and competitive differentiation of trends, leading to irrelevant, risky, generic, and untimely content that does not align with the publisher's message or expertise.

Method used

A content generation device that normalizes trend signals into a common schema, evaluates their lifecycle and risk, generates content guidelines based on a publisher's knowledge base, and verifies the content for differentiation and safety before publication.

Benefits of technology

Generates safe, differentiated, and timely content by excluding irrelevant or risky trends, ensuring alignment with the publisher's message and expertise, and optimizing publication timing based on trend lifecycle stages.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure R1020260132511_ABST
    Figure R1020260132511_ABST
Patent Text Reader

Abstract

A content generation device according to one embodiment of the present invention comprises: a trend normalization unit that converts trend signals collected from a plurality of sources into a common schema including the meaning of the trend and the amount of temporal change; a trend evaluation unit that calculates usability based on a lifecycle and risk level based on the amount of temporal change; a guideline generation unit that generates content generation guidelines for trends in which the usability meets criteria; and a content synthesis unit that generates candidate content based on the content generation guidelines, wherein the content generation guidelines are guidelines that connect the trends with assets of a target knowledge base. Accordingly, risky or irrelevant trends can be excluded before content generation, and differentiated content reflecting the unique assets of the target can be generated.
Need to check novelty before this filing date? Find Prior Art

Description

Technology Field

[0001] The present invention relates to a content generation technology, and more specifically, to a method and apparatus for generating content by evaluating a trend signal and connecting it to a target knowledge base. Background Technology

[0003] When content publishers (e.g., brands, individual creators, media outlets, or organizations) post content through channels such as social media, content related to topics that are currently attracting a lot of attention tends to receive higher exposure and engagement compared to content that is not. Accordingly, attempts are being made to utilize trending topics, search terms, memes, industry issues, or conversation flows in content creation.

[0004] Conventional methods for generating content utilizing trends generally remained limited to inserting popular search terms or trend keywords into content creation prompts, or using human-selected trends as content material.

[0005] However, this method has the following problems.

[0006] First, since only the superficial keywords of a trend are utilized without considering its meaning or context, trends that are substantially irrelevant to the publisher may be used in content, or content in which trends are unnaturally inserted may be generated.

[0007] Second, even if a trend includes sensitive or risky topics, failing to identify them in advance can lead to the utilization of the trend actually damaging the publisher's reputation.

[0008] Third, as multiple publishers utilizing trends repeat similar hooking methods, the generated content may fail to differentiate itself from competing content and remain generic.

[0009] Fourth, if content reflecting trends is not connected to the publisher's core message, product, or expertise, interest in the trend may not lead to interest in the publisher.

[0010] Fifth, although trends have a lifecycle in which interest changes over time, the effectiveness of utilizing trends may be limited because the lifecycle of these trends or the suitability of the timing of publication are not considered.

[0011] Therefore, technology is required to evaluate trends not merely as surface keywords, but from the perspectives of their meaning, risk, lifecycle, competitive congestion, and potential for connection with publishers; to generate content by connecting selected trends to the publisher's knowledge base based on the evaluation results; and to verify the generated content according to the publisher's policies and differentiation criteria. Prior art literature

[0013] Korean Registered Patent No. 10-2990127 (July 8, 2026), AI-based brand matching and content generation method The problem to be solved

[0014] The objective of the present invention is to provide a method and apparatus for generating trend-based content that evaluates trends in terms of their meaning, risk, lifecycle, and potential for connection with a publishing entity, and generates safe and differentiated content by connecting selected trends based on the evaluation results to a knowledge base of a publishing entity.

[0015] Another objective of the present invention is to provide a trend-based content generation method and apparatus that normalizes trend signals into a common trend schema and calculates the lifecycle stage and availability of the trend to exclude trends that are irrelevant to or dangerous to the publisher before content generation.

[0016] Another objective of the present invention is to provide a trend-based content generation method and apparatus that generates content differentiated from competitive content by calculating a differentiation score based on similarity with competitive content for a generated content angle, and verifying the content based on the differentiation score and the policy of the publishing entity.

[0017] However, the technical problems that the present invention aims to solve are not limited to those described above, and other unmentioned problems can be clearly understood by those skilled in the art from the description of the invention below. means of solving the problem

[0019] According to one embodiment of the present invention to achieve the above objective, a content generation device is disclosed, comprising: a trend normalization unit that converts a trend signal collected from a plurality of sources into a common schema including the meaning of the trend and the amount of temporal change; a trend evaluation unit that calculates usability for the converted trend based on a lifecycle and risk level based on the amount of temporal change; a guideline generation unit that generates content generation guidelines for a trend in which the usability meets criteria; and a content synthesis unit that generates candidate content based on the content generation guidelines, wherein the content generation guidelines are guidelines that connect the trend with an asset of a target knowledge base.

[0020] In one embodiment, the content generating device may further include a verification unit that verifies whether the candidate content satisfies preset conditions and classifies the candidate content as content awaiting publication or unpublished content according to the verification result.

[0021] In one embodiment, the verification unit calculates a differentiation score based on the similarity with reference content for the candidate content, and the preset condition may include the differentiation score being greater than or equal to a standard.

[0022] In one embodiment, the differentiation score may be calculated based on at least one of string similarity for a title or introductory phrase between the candidate content and the reference content, n-gram redundancy, and embedding similarity in a representation space.

[0023] In one embodiment, the verification unit verifies the factual accuracy of the candidate content or whether there is a misunderstanding of the trend context using at least one of the trend source text, news summary, internal knowledge base, and approved source documents, and the verification result may be further used as the preset condition.

[0024] In one embodiment, the differentiation score may be calculated by further considering the degree to which the assets of the target knowledge base are reflected in the candidate content.

[0025] In one embodiment, the trend evaluation unit can calculate the usability by combining two or more of the relevance score with the target knowledge base, the growth rate score, the lifecycle timing score, the sensitivity score, the legal risk score, and the competitive congestion score.

[0026] In one embodiment, the trend evaluation unit can classify the lifecycle stage of the trend using two or more of the current interest, rising speed, rising acceleration, and observation duration from the hourly observations of the trend signal.

[0027] In one embodiment, the content generating device may further include a feedback unit that records the performance of the published content and adjusts at least one of the criteria for calculating the usability, the criteria for calculating the differentiation score, and the publishing timing for each lifecycle stage based on the performance.

[0028] In one embodiment, the content synthesis unit may further calculate a publishing time score of the candidate content based on the lifecycle stage of the trend.

[0029] According to one embodiment of the present invention for achieving the above objective, a content generation method performed by a content generation device comprises: a step of converting a trend signal collected from a plurality of sources into a common schema including the meaning of the trend and the amount of temporal change; a step of calculating usability for the converted trend based on a lifecycle and risk based on the amount of temporal change; a step of generating content generation guidelines for a trend in which the usability meets criteria; and a step of generating candidate content based on the content generation guidelines, wherein the content generation guidelines are guidelines that connect the trend with an asset of a target knowledge base.

[0030] In one embodiment, the content generation method may further include a step of verifying whether the candidate content satisfies a preset condition; and a step of classifying the candidate content into content awaiting publication or unpublished content according to the verification result.

[0031] In one embodiment, the verification step includes the step of calculating a differentiation score based on the similarity between the candidate content and the reference content, and the preset condition may include the differentiation score being greater than or equal to a standard.

[0032] In one embodiment, the differentiation score may be calculated based on at least one of string similarity for a title or introductory phrase between the candidate content and the reference content, n-gram redundancy, and embedding similarity in a representation space.

[0033] In one embodiment, the verification step includes verifying the factual accuracy of the candidate content or whether there is a misunderstanding of the trend context using at least one of the original trend text, news summary, internal knowledge base, and approved source documents, and the verification result may be further used as the preset condition.

[0034] In one embodiment, the differentiation score may be calculated by further considering the degree to which the assets of the target knowledge base are reflected in the candidate content.

[0035] In one embodiment, the step of calculating the usability may include the step of calculating the usability by combining two or more of the relevance score with the target knowledge base, the growth rate score, the lifecycle timing score, the sensitivity score, the legal risk score, and the competitive congestion score.

[0036] In one embodiment, the step of calculating the usability may include classifying the lifecycle stage of the trend using two or more of the current interest, rising speed, rising acceleration, and observation duration from the hourly observations of the trend signal.

[0037] In one embodiment, the content generation method may further include the step of recording the performance of the published content; and the step of adjusting at least one of the criteria for calculating the usability, the criteria for calculating the differentiation score, and the publishing timing for each lifecycle stage based on the performance.

[0038] In one embodiment, the step of generating the candidate content may further include the step of calculating a publication time score of the candidate content based on the lifecycle stage of the trend. Effects of the invention

[0040] According to one embodiment of the present invention, by normalizing a trend signal into a common schema including its meaning and temporal change amount, calculating usability based on the trend's lifecycle and risk level, and then proceeding with content creation only for trends that satisfy these criteria, it is possible to exclude trends that are irrelevant to the target or dangerous from being utilized in content at a stage prior to content creation. Accordingly, unlike conventional methods that insert only superficial keywords, content that is safely utilized by considering the context and risk level of the trend can be generated.

[0041] According to one embodiment of the present invention, a differentiation score is calculated for generated candidate content based on the similarity to reference content and the degree to which assets of the target knowledge base are reflected, and this score is used as a condition for publication. This allows for the selective publication of differentiated content that reflects the unique perspective of the target, rather than remaining at the level of general expressions similar to competing content. Furthermore, by providing a dual gating structure that determines the usability of the trend itself before content creation and verifies the factual accuracy of the generated candidate content and whether there is a misunderstanding of the trend context after content creation, the risks associated with utilizing trends can be managed in two stages.

[0042] According to one embodiment of the present invention, the lifecycle stage of a trend is classified using temporal change amounts, including rising speed and rising acceleration, from hourly observations of a trend signal, and the publishing time of candidate content is determined based on the lifecycle stage, thereby allowing content to be published at a time that aligns with the flow of changing interest in the trend. Furthermore, by recording the performance of the published content and adjusting the criteria for calculating usability, the criteria for calculating differentiation scores, and the publishing timing for each lifecycle stage, the results of trend utilization can be reflected in the subsequent content creation strategy to enable continuous improvement.

[0043] However, the effects obtainable through the present invention are not limited to those described above, and other unmentioned technical effects will be clearly understood by those skilled in the art from the description of the invention below. Brief explanation of the drawing

[0045] FIG. 1 is a block diagram schematically showing the configuration of a content creation device according to one embodiment of the present invention. FIG. 2 is a flowchart showing the overall flow of a content creation method according to one embodiment of the present invention. FIG. 3 is a flowchart illustrating the process of calculating the usability of a trend according to one embodiment of the present invention. FIG. 4 is a flowchart illustrating the process of verifying candidate content according to one embodiment of the present invention. FIG. 5 is a diagram showing the relationship between the lifecycle stage of a trend and the time of publication according to one embodiment of the present invention. FIG. 6 is a block diagram showing the hardware configuration of a content creation device according to one embodiment of the present invention. Specific details for implementing the invention

[0046] An embodiment of the present invention will be described in detail below with reference to the attached drawings. Prior to this, terms and words used in this specification and claims should not be interpreted as being limited to their ordinary or dictionary meanings. Instead, based on the principle that the inventor can appropriately define the concept of a term to best describe their invention, they should be interpreted in a meaning and concept consistent with the technical spirit of the present invention. Therefore, the embodiments described in this specification and the configurations illustrated in the drawings are merely partial embodiments of the present invention and do not represent all of the technical spirit of the present invention. It should be understood that various equivalents and modifications capable of replacing them may exist at the time of filing this application.

[0047] Additionally, as used herein, “comprise, include” and / or “comprising, including” specify the presence of the mentioned features, numbers, steps, actions, parts, elements, and / or groups thereof, and do not exclude the presence or addition of one or more other features, numbers, actions, parts, elements, and / or groups.

[0048] Additionally, to aid in understanding the invention, the attached drawings are not drawn to actual scale, and the dimensions of some components may be exaggerated. Furthermore, the same reference numerals may be assigned to identical components in different embodiments.

[0049] The statement that two subjects of comparison are 'identical' means that they are 'substantially identical.' Therefore, substantial identity may include deviations considered low in the industry, for example, deviations within 5%. Additionally, the statement that a parameter is uniform in a given area may mean that it is uniform from an average perspective.

[0050] Although terms such as "first," "second," etc., are used to describe various components, it goes without saying that these components are not limited by these terms. These terms are used merely to distinguish one component from another, and unless specifically stated otherwise, the first component may also be the second component.

[0051] Throughout the specification, unless specifically stated otherwise, each component may be singular or plural.

[0052] The fact that any configuration is placed on the "upper (or lower)" of a component or on the "upper (or lower)" of a component may mean not only that any configuration is placed in contact with the upper (or lower) surface of said component, but also that another configuration may be interposed between said component and any configuration placed on (or below) said component.

[0053] Furthermore, where one component is described as being "on," "connected to," or "coupled to" another component, it should be understood that while the components may be directly connected or coupled to each other, another component may be "interposed" between each component, or that each component may be "connected," "coupled," or "coupled" through another component.

[0054] As used herein, the term “and / or” includes any and all combinations of one or more of the associated listed items. Additionally, when describing embodiments of the invention, the use of “may” relates to “one or more embodiments of the invention.” Expressions such as “one or more” preceding a list of elements modify the entire list of elements and do not modify individual elements of the list.

[0055] Throughout the specification, "A and / or B" means A, B, or A and B unless specifically stated otherwise, and "C to D" means C or more and D or less, unless specifically stated otherwise.

[0056] When syntax such as "at least one of A, B, and C", "at least one of A, B, or C", "at least one selected from the group of A, B, and C", or "at least one selected from A, B, and C" is used to specify a list of elements A, B, and C, the syntax can refer to any suitable combination.

[0057] The term "use" may be considered synonymous with the term "utilize." As used herein, terms such as "substantially," "about," and similar terms are used as terms of approximation rather than degree, and are intended to account for the inherent variation of a measured or calculated value that would be recognized by a person skilled in the art.

[0058] In this specification, terms such as first, second, third, etc. may be used to describe various elements, components, regions, layers, and / or sections, but these elements, components, regions, layers, and / or sections should not be limited by these terms. These terms are used to distinguish one element, component, region, layer, or section from another element, component, region, layer, or section. Accordingly, the first element, component, region, layer, or section discussed below may be named the second element, component, region, layer, or section without departing from the teachings of the exemplary embodiments.

[0059] Spatial relative terms such as "beneath," "below," "lower," "above," and "upper" may be used in the specification for ease of description to explain the relationship between one element or feature and other element(s) or feature(s) as illustrated in the drawings. Spatially relative positions are to be understood as encompassing different orientations of the device in use or operation, in addition to the orientations depicted in the drawings. For example, if the device in the drawings is inverted, the other element is understood as "below" or "below," and the described element is understood as "above" or "upper" of the other element. Thus, the term "below" may encompass both the up and down directions.

[0060] The terms used in this specification are intended to describe embodiments of the invention and are not intended to limit the invention.

[0061] In this specification, the term "trend" refers to a topic, expression, search query, hashtag, conversation flow, or content pattern observed from multiple signal sources, which may refer to a change in collective interest in the said topic over time. A trend may be characterized by at least semantics and temporal change, and may be expressed by further including one or more of interest, lifecycle stage, sensitivity, and competitive congestion.

[0062] In this specification, the term "trend signal" may refer to data observed from one or more signal sources indicating the existence of a specific trend or a change therein. Trend signals may include, but are not limited to, one or more of search terms, post topics, hashtags, news headlines, community conversations, industry keywords, meme signals, trending queries, or competitive content patterns.

[0063] In this specification, the term "target knowledge base" may refer to a set of data that stores information related to an entity publishing content. The entity publishing content may be an identifiable entity, such as a brand, individual, organization, or media, and is not limited to a specific entity. The target knowledge base may include one or more of the entity's messages, product features, customer issues, industry expertise, prohibited topics, tone guides, and past content performance. Individual information stored in the target knowledge base and used for content creation may be referred to as an "asset."

[0064] In this specification, "content generation instruction" refers to data that defines the connection relationship between a trend and an asset of a target knowledge base, and may refer to instructions indicating which asset should be associated with the trend to generate content from which perspective. The content generation instruction may be used as input for generating candidate content and may also be referred to as a connection angle in one embodiment.

[0065] In this specification, "reference content" refers to content that serves as a subject of comparison for evaluating the degree of differentiation of candidate content, and may refer to existing content related to the trend. Reference content may include content dealing with the same or similar trends, content of competitors, or generally generated content.

[0066] FIG. 1 is a block diagram schematically showing the configuration of a content creation device according to one embodiment of the present invention.

[0067] The content generation device (100) is a device that generates content by evaluating trends and connecting them to a target knowledge base, and can collect and normalize trend signals, calculate the usability of the normalized trends, generate content generation guidelines for trends determined to be usable, and generate candidate content based on the content generation guidelines. The content generation device (100) can be implemented as a server, a computing device, or a combination thereof, and is not limited to a specific form.

[0068] Referring to FIG. 1, the content creation device (100) may include one or more of a trend collection unit (110), a trend normalization unit (120), a trend evaluation unit (130), a guideline generation unit (140), a content synthesis unit (150), a verification unit (160), and a feedback unit (170). Some of the above components may be omitted, and the content creation device (100) may include additional components other than those above. Each of the above components may be a functional configuration implemented by a processor described later executing instructions stored in memory, and does not necessarily mean a physically distinct configuration.

[0069] The trend collection unit (110) can collect trend signals from multiple sources. The multiple sources may include one or more of search term rising data, social media popular topics, news headlines and article summaries, community posts and comments, industry reports or technical blogs, patterns of public content from competitors, or keywords that are rising in internal customer inquiries or sales conversations. In one embodiment, the collected trend signals may be stored along with one or more of time, source, region, language, interest, and source context. The trend signals may be collected from real-time streams, batch collection, manual input, or from external data providers. Meanwhile, the trend collection unit (110) is configured according to one embodiment, whereas in another embodiment, the trend signals may be received from an external device and provided directly to the trend normalization unit (120).

[0070] The trend normalization unit (120) can convert the trend signal collected by the trend collection unit (110) into a common schema including the meaning of the trend and the amount of temporal change. Trend signals from different sources can be expressed in a unified structure by being converted into the common schema. In one embodiment, the common schema may be composed of fields including one or more of a trend identifier, surface representation, semantic summary, signal source type, rise rate, interest level, lifecycle stage, sensitivity, competitive congestion, and target relevance. Here, the amount of temporal change may include the rise rate and may further include the rise acceleration.

[0071] The trend evaluation unit (130) can calculate usability based on the lifecycle and risk level based on the amount of temporal change for the trend transformed by the trend normalization unit (120). The trend evaluation unit (130) can calculate usability to select trends that the subject can use safely and naturally, rather than simply trends with high interest. In one embodiment, the trend evaluation unit (130) may use one or more of the following as evaluation factors: current interest, rate of rise, lifecycle stage, subject relevance, sensitive topic risk, legal or regulatory risk, competitive congestion, differentiation potential, and content creation potential. The trend evaluation unit (130) may perform the evaluation using one or more of the following: rule-based evaluation, embedding similarity, generative artificial intelligence evaluation, or classification model. The specific operation of the trend evaluation unit (130) will be described later with reference to FIG. 3.

[0072] The guideline generation unit (140) can generate content generation guidelines for trends whose usability, calculated by the trend evaluation unit (130), meets the criteria. The content generation guidelines may be guidelines that connect the trend with the assets of the target knowledge base. In one embodiment, the guideline generation unit (140) can generate connectable content angles by comparing the meaning of the trend with the target knowledge base, and the content angles may include one or more of the following: an angle that explains customer problems revealed by the trend using the expertise of the target, an angle that corrects misunderstandings of the trend, an angle that connects the trend with product use cases, or an angle that converts questions arising from the trend into educational content. In one embodiment, the guideline generation unit (140) can convert the meaning of the trend and the assets of the target knowledge base into a common expression space and select the assets based on their relevance in the common expression space.

[0073] The content synthesis unit (150) can generate candidate content based on content generation guidelines generated by the guideline generation unit (140). In one embodiment, the candidate content may be generated in one or more formats such as a short social post, a long descriptive post, a comment or reply, an image or video description, a campaign phrase, a newsletter summary, or a community Q&A. The content synthesis unit (150) can generate candidate content using a prompt that includes a description of the connection between the trend and the target message so that the trend is not inserted unnaturally. In one embodiment, the content synthesis unit (150) can further calculate a posting point score of the candidate content based on the lifecycle stage of the trend.

[0074] The verification unit (160) verifies whether candidate content generated by the content synthesis unit (150) satisfies preset conditions, and classifies the candidate content as content awaiting publication or unpublished content according to the verification result. In one embodiment, the verification unit (160) may calculate a differentiation score based on similarity with reference content for the candidate content, and the preset conditions may include the differentiation score being greater than or equal to a standard. In one embodiment, the verification unit (160) may verify the factual accuracy or misunderstanding of the trend context of the candidate content using one or more of the trend source text, news summary, internal knowledge base, or approved source documents, and may further use the verification result as the preset conditions. In one embodiment, the verification unit (160) may examine one or more of the target tone suitability, misunderstanding of the trend context, sensitive topics or legal risks, excessive promotional content, prohibited words or topics, differentiation score criteria, factual accuracy, and suitability for the time of publication. The candidate content classified as unpublished content may be classified into one of modification, human approval, disposal, or regeneration. The specific operation of the verification unit (160) will be described later with reference to FIG. 4.

[0075] The feedback unit (170) records the performance of the posted content and, based on the performance, can adjust one or more of the criteria for calculating the usability, the criteria for calculating the differentiation score, and the posting timing for each lifecycle stage. In one embodiment, the performance recorded by the feedback unit (170) can be used for one or more of adjusting the target relevance criteria, adjusting the differentiation score criteria, learning the posting timing for each trend lifecycle, generating avoidance rules for a specific trend type, or storing high-performance connection angle templates. Accordingly, the content creation device (100) can reflect the results of trend utilization in the subsequent content creation strategy.

[0076] FIG. 2 is a flowchart showing the overall flow of a content generation method according to an embodiment of the present invention. Each step described below can be performed by the content generation device (100) of FIG. 1.

[0077] First, the content creation device (100) can collect trend signals from multiple sources (S210).

[0078] The above trend signals may be collected from one or more of search term rise data, social media popular topics, news headlines, community conversations, industry reports, competitive content patterns, or internal customer inquiries.

[0079] Next, the content generation device (100) can convert the collected trend signal into a common schema including the meaning of the trend and the amount of temporal change (S220).

[0080] Trend signals from different sources can be expressed in a unified structure by being converted into the above common schema.

[0081] The content creation device (100) can calculate usability based on the lifecycle and risk level based on the amount of temporal change for the converted trend (S230).

[0082] The above usability may indicate the extent to which the trend can be used safely and naturally by the target. The specific process for calculating usability will be described later with reference to FIG. 3.

[0083] The content creation device (100) can determine whether the calculated usability meets the criteria (S240).

[0084] If it is determined that the above usability does not meet the criteria, the content creation device (100) may exclude the trend and store risk information (S241). Information regarding the excluded trend in this way may be used to evaluate subsequent trends or to create avoidance rules.

[0085] Meanwhile, if the above-mentioned usability is determined to meet the criteria, the content creation device (100) can determine whether the risk level of the trend falls within a preset intermediate range (S245).

[0086] In one embodiment, if the risk level is determined to be in the intermediate range, the content creation device (100) may classify the trend as a target for human approval or enhanced verification (S246). That is, the content creation device (100) may exclude trends with usability below a standard and classify trends with usability above a standard and a risk level in the intermediate range as targets for human approval or enhanced verification, thereby processing the trends differentially according to the risk level.

[0087] If the above risk level is determined not to fall within the above intermediate range, or if the above human approval is granted, the content creation device (100) can generate content creation guidelines for the corresponding trend that connect the trend with the assets of the target knowledge base (S250).

[0088] The content creation device (100) can create candidate content based on the content creation guidelines (S260).

[0089] Next, the content generation device (100) can verify whether the candidate content satisfies a preset condition (S270). The specific process of the verification is described later with reference to FIG. 4.

[0090] The content creation device (100) can determine whether the candidate content satisfies the preset conditions (S280).

[0091] If it is determined that the above candidate content does not satisfy the above preset conditions, the content creation device (100) may classify the above candidate content as unpublished content, and the unpublished content may be classified into one of modification, human approval, disposal, or regeneration (S281). In one embodiment, if the above candidate content is classified as regeneration, the above candidate content may be regenerated by returning to the step of creating candidate content based on content creation guidelines (S260). In another embodiment, the regeneration may be performed by returning to the step of creating content creation guidelines (S250).

[0092] Meanwhile, if the above candidate content is determined to satisfy the above preset conditions, the content creation device (100) may classify the above candidate content as content awaiting publication and then publish it (S290).

[0093] Subsequently, the content creation device (100) records the performance of the posted content and, based on the performance, can adjust one or more of the criteria for calculating the usability, the criteria for calculating the differentiation score, and the posting timing for each lifecycle stage (S295). Accordingly, the content creation device (100) can reflect the results of trend utilization in the subsequent content creation.

[0094] Each of the above steps does not necessarily have to be performed in the order described, and some steps may be omitted or performed in a different order, and other steps may be added.

[0095] FIG. 3 is a flowchart illustrating a process for calculating the usability of a trend according to an embodiment of the present invention. Each step of FIG. 3 can be performed by the trend evaluation unit (130) of FIG. 1 and can be included in the step (S230) of calculating the usability of FIG. 2.

[0096] The trend evaluation unit (130) can obtain hourly observations of the trend signal (S310).

[0097] The above hourly observations may include multiple values ​​observed over time for a single trend, and may include, for example, one or more of hourly interest, hourly mention volume, or hourly search volume.

[0098] The trend evaluation unit (130) can calculate the current interest, rising speed, rising acceleration, and observation duration from the above hourly observation values ​​(S320). Here, the rising speed may correspond to the amount of change in interest compared to the previous period, the rising acceleration may correspond to the amount of change in the rising speed, and the observation duration may correspond to the elapsed time from the point when the trend began to be observed. The current interest, rising speed, rising acceleration, and observation duration may be values ​​representing the amount of temporal change of the trend.

[0099] The trend evaluation unit (130) can classify the lifecycle stage of the trend using two or more of the current interest, rising speed, rising acceleration, and observation duration (S330).

[0100] In one embodiment, the lifecycle stage may be classified into any one of initial, rising, peak, and falling. For example, the trend evaluation unit (130) may classify the trend into an initial or rising stage when interest is low and the rising speed begins to increase to a positive (+) value, classify it into a peak stage when the rising acceleration approaches zero and interest reaches a maximum, and classify it into a falling stage when interest decreases. In this way, the trend evaluation unit (130) can distinguish between a trend in the rising phase and a trend in the falling phase, even if the current interest is the same, by utilizing not only the interest of the trend itself but also the speed and acceleration of its change.

[0101] The trend evaluation unit (130) can calculate one or more of the following: a relevance score with the target knowledge base, a rise speed score, a lifecycle timing score, a sensitivity score, a legal risk score, and a competitive congestion score (S340).

[0102] In one embodiment, the relevance score may be calculated by converting the semantic summary of the trend and the assets of the target knowledge base into a common representation space, and the lifecycle timing score may be calculated based on the classified lifecycle stage. The sensitivity score and the legal risk score may each indicate the degree to which the trend includes sensitive topics or legal and regulatory risks, and the competitive congestion may indicate the degree of density of competing content using the same trend.

[0103] The trend evaluation unit (130) can calculate usability by combining two or more of the above-calculated scores (S350).

[0104] In one embodiment, the trend evaluation unit (130) may calculate the usability by adding one or more of the relevance score, growth rate score, lifecycle timing score, and differentiation possibility with the target knowledge base, and subtracting one or more of the sensitivity score, legal risk score, and competitive congestion. For example, the usability may be calculated according to the following relationship.

[0105] Availability = Target Relevance Score + Rate of Ascent Score + Lifecycle Timing Score + Differentiability - Sensitivity Score - Legal Risk Score - Competition Congestion

[0106] In one embodiment, weights may be assigned to each score when calculating the usability, and each score may be normalized to a preset range. The weights may be adjusted based on performance recorded by the feedback unit (170) described later.

[0107] The usability calculated in this way can be compared with a standard in step (S240) of FIG. 2, and trends where the usability is below the standard are excluded, while trends where the usability is above the standard can be the target for creating content creation guidelines.

[0108] FIG. 4 is a flowchart illustrating a process for verifying candidate content according to an embodiment of the present invention. Each step of FIG. 4 can be performed by the verification unit (160) of FIG. 1 and can correspond to the verification step (S270) and the determination step (S280) of FIG. 2.

[0109] The verification unit (160) can receive candidate content and reference content (S410).

[0110] The above candidate content may be content generated by the content synthesis unit (150) of FIG. 1, and the above reference content may be existing content related to the trend, and may include one or more of content dealing with the same or similar trend, content of a competitor, or generally generated content.

[0111] The verification unit (160) can calculate the similarity between the candidate content and the reference content (S420).

[0112] In one embodiment, the similarity may be calculated based on one or more of string similarity, n-gram redundancy, and embedding similarity in a representation space for the title or introductory phrase of the candidate content and the reference content. Here, the string similarity may represent the degree of matching of strings, the n-gram redundancy may represent the degree of redundancy of a subsequence consisting of n consecutive tokens, and the embedding similarity may represent the similarity between vectors calculated by converting the candidate content and the reference content into vectors in a representation space, respectively.

[0113] The verification unit (160) can calculate the extent to which the assets of the target knowledge base are reflected in the candidate content (S430).

[0114] In one embodiment, the degree to which the asset is reflected may be calculated based on the extent to which one or more of the target's product name, unique expression, or unique perspective are included in the candidate content, for example, based on the frequency of appearance of the asset or the number of elements matching the asset.

[0115] The verification unit (160) can calculate a differentiation score based on the similarity and the degree to which the asset is reflected (S440).

[0116] In one embodiment, the differentiation score may be calculated such that it has a higher value as the similarity with the reference content is lower, and a higher value as the degree to which the asset is reflected is higher. That is, the differentiation score may be a value that reflects both how different the candidate content is from the competing content and how much the target's unique asset is reflected. In one embodiment, the verification unit (160) may assign a low differentiation score to candidate content where the similarity is above a standard and the degree to which the asset is reflected is low.

[0117] The verification unit (160) can verify the factual accuracy of the candidate content and whether there is a misunderstanding of the trend context using the source document (S450).

[0118] In one embodiment, the source document may include one or more of the original trend text, news summary, internal knowledge base, or approved source document, and the verification unit (160) may verify whether the candidate content includes claims that do not match the source document or misinterpret the context of the trend.

[0119] The verification unit (160) can determine whether the differentiation score is above the standard and whether the factuality is satisfied (S460).

[0120] In one embodiment, the preset condition may include the differentiation score being greater than or equal to a standard, and the verification result regarding whether there is a misunderstanding of the factuality or trend context may be further used as the preset condition.

[0121] If the above differentiation score is above the standard and the above factuality is determined to be satisfied, the verification unit (160) can classify the above candidate content as content awaiting posting (S470).

[0122] Meanwhile, if the above differentiation score is below the standard or the above factuality is determined not to be satisfied, the verification unit (160) may classify the above candidate content as unpublished content, and the above unpublished content may be classified into one of modification, human approval, disposal, or regeneration (S461).

[0123] In one embodiment, the candidate content classified as regenerated among the unpublished content may be regenerated by returning to the step of generating candidate content (S260) or the step of generating content generation guidelines (S250) as described with reference to FIG. 2.

[0124] In one embodiment, the verification unit (160) may use one or more of the target tone suitability, sensitive subject or legal risk, excessive promotional nature, prohibited words or prohibited subject and time of publication suitability as additional preset conditions in addition to the differentiation score and the factuality.

[0125] FIG. 5 is a diagram showing the relationship between the lifecycle stage of a trend and the time of publication according to an embodiment of the present invention. In FIG. 5, the horizontal axis represents time, and the vertical axis represents the level of interest in the trend.

[0126] Referring to FIG. 5, interest in a trend may change over time, and this change in interest may be divided into lifecycle stages of initial, rising, peak, and falling. In the initial stage, interest may remain at a low level, in the rising stage, interest may increase, in the peak stage, interest may reach a maximum, and in the falling stage, interest may decrease. As previously explained with reference to FIG. 3, the lifecycle stages may be classified by the trend evaluation unit (130) using two or more of the current interest, rising speed, rising acceleration, and observation duration.

[0127] The content synthesis unit (150) can calculate a publication time score of candidate content based on the lifecycle stage of the trend. The publication time score may indicate the degree of suitability for publishing content utilizing the trend. In one embodiment, the publication time score may be calculated to have a relatively high value when the trend is in the rising stage, a medium value when it is in the peak stage, and a relatively low value when it is in the falling stage.

[0128] For example, since the point in time when the trend is in the rising phase is a time when interest has not yet reached its peak and there is room for further increase, while sufficient interest has already been formed, the content synthesis unit (150) can determine that the point in time belonging to the rising phase is a suitable time for posting. On the other hand, since the point in time when the trend has already passed its peak and is in the falling phase is a time when interest is decreasing and exposure and reaction to the content may be limited, the content synthesis unit (150) can assign a low posting time score to the point in time belonging to the falling phase.

[0129] In this way, the content creation device (100) considers not only the current interest level of the trend but also the lifecycle stage of the trend, thereby distinguishing between the rising phase and the falling phase even if the interest level is the same, and determining the posting time. The above posting time score can be used to determine whether to post the candidate content or the order of posting.

[0130] In one embodiment, the suitability of the posting time for each lifecycle stage may be adjusted based on performance recorded by the feedback unit (170) described below. That is, as the performance of the content posted at a specific lifecycle stage is recorded, the content creation device (100) may adjust the posting time score assigned to each lifecycle stage.

[0131] FIG. 6 is a block diagram showing the hardware configuration of a content creation device according to an embodiment of the present invention. The content creation device (900) of FIG. 6 may correspond to the content creation device (100) of FIG. 1.

[0132] Referring to FIG. 6, the content creation device (900) may include one or more of a memory (910), a communication interface (920), and a processor (930). Some of the above components may be omitted, and the content creation device (900) may include additional components in addition to the above components.

[0133] The memory (910) may store one or more instructions regarding the content creation device (900). The memory (910) may store an operating system and various software programs for operating the content creation device (900), and may store data for the content creation device (900) to operate. In one embodiment, the memory (910) may store one or more of a target knowledge base, collected trend signals, trends converted into a common schema, content creation guidelines, candidate content, and publication performance. The memory (910) may include one or more of semiconductor memory such as flash memory or magnetic storage media such as a hard disk, but is not limited thereto.

[0134] The communication interface (920) is configured to communicate with an external device and can perform communication with the external device based on a wired or wireless communication method. In one embodiment, the communication interface (920) communicates with one or more of a source providing a trend signal, an external device storing a target knowledge base, or a publishing platform where content is published, so that the content creation device (900) can transmit and receive data necessary for creating and providing content.

[0135] The processor (930) can control the overall operation of the content creation device (900). The processor (930) can control the operation of the content creation device (900) by executing one or more instructions stored in memory (910). In one embodiment, the processor (930) can perform one or more operations of the trend collection unit (110), trend normalization unit (120), trend evaluation unit (130), guideline generation unit (140), content synthesis unit (150), verification unit (160), and feedback unit (170) described with reference to FIG. 1 by executing instructions stored in memory (910). That is, each component of FIG. 1 may be a functional configuration implemented by the processor (930) executing instructions stored in memory (910), and may be implemented by a single processor (930) or distributed among multiple processors.

[0136] In one embodiment, each step described with reference to FIGS. 2 through 4 may be performed by the processor (930) executing an instruction stored in memory (910). Additionally, the instruction that causes each step to be performed may be provided by being stored in a computer-readable recording medium. The computer-readable recording medium may be provided in the form of a non-transitory storage medium, where a non-transitory storage medium may mean a storage medium that exists without containing a signal.

[0137] As described above, according to one embodiment of the present invention, the content generation device (100) excludes dangerous or irrelevant trends by evaluating trends based on lifecycle and risk level before content generation, selects generated candidate content based on a differentiation score based on similarity with reference content and the degree of asset reflection of the target knowledge base, determines the publication time according to the lifecycle stage of the trend, and reflects the publication performance in subsequent content generation, thereby enabling the generation of content that safely and differentially utilizes trends at a timely moment.

[0138] Meanwhile, the method according to the present invention is written as a program executable on a computer and can be implemented on various recording media such as magnetic storage media, optical reading media, and digital storage media.

[0139] Implementations of the various technologies described herein may be implemented in digital electronic circuits, computer hardware, firmware, software, or a combination thereof. Such implementations may be implemented as computer programs stored on a computer-readable storage medium to control the operation of a data processing device. Such implementations may be implemented as computer program products, i.e., information carriers, such as computer programs stored on a machine-readable storage device or a computer-readable non-transient storage medium, for processing by the operation of a data processing device, e.g., a programmable processor, a computer, or a plurality of computers, or for controlling such operation. Computer programs such as the computer program(s) described above may be written in any form of programming language including compiled or interpreted languages, and may be deployed in any form including as a standalone program or as a module, component, subroutine, or other unit suitable for use in a computing environment. Computer programs may be deployed to be processed on one computer or a plurality of computers at a single site, or distributed across a plurality of sites and interconnected by a communication network.

[0140] Processors suitable for processing computer programs include, for example, both general-purpose and special-purpose microprocessors, and any one or more processors of any type of digital computer. Generally, the processor will receive instructions and data from read-only memory or random access memory, or both. The elements of the computer may include at least one processor that executes instructions and one or more memory devices that store instructions and data. Generally, the computer may include one or more mass storage devices that store data, for example, magnetic, magneto-optical disks, or optical disks, or may be combined to receive data from these, transmit data to these, or both. Information carriers suitable for embodying computer program instructions and data include, for example, semiconductor memory devices, magnetic media such as hard disks, floppy disks, and magnetic tapes; optical media such as CD-ROMs (Compact Disk Read Only Memory) and DVDs (Digital Video Disks); magneto-optical media such as floptical disks; ROMs (Read Only Memory); RAMs (Random Access Memory); flash memory; EPROMs (Erasable Programmable ROM); EEPROMs (Electrically Erasable Programmable ROM); etc. Processors and memory may be supplemented by or included in special-purpose logic circuit organizations.

[0141] Although this specification contains details of a number of specific embodiments, they should not be understood as limiting the scope of any invention or claimables, but rather as descriptions of features that may be characteristic of a specific embodiment of a specific invention. Specific features described in this specification in the context of individual embodiments may be implemented in combination in a single embodiment. Conversely, various features described in the context of a single embodiment may also be implemented in multiple embodiments individually or in any appropriate sub-combination. Furthermore, while features may operate in a specific combination and be described as initially claimed, one or more features from the claimed combination may be excluded from the combination in some cases, and the claimed combination may be changed to a sub-combination or a variation of the sub-combination.

[0142] Likewise, although operations are depicted in the drawings in a specific order, this should not be understood as requiring that such operations be performed in that specific or sequential order depicted to obtain a desirable result, or that all depicted operations must be performed. In certain cases, multitasking and parallel processing may be advantageous. Furthermore, the separation of the various device components of the above-described embodiments should not be understood as requiring such separation in all embodiments, and it should be understood that the described program components and devices can generally be integrated together into a single software product or packaged into multiple software products.

[0143] Meanwhile, the embodiments of the present invention disclosed in this specification and drawings are merely specific examples provided to aid understanding and are not intended to limit the scope of the present invention. It is obvious to those skilled in the art that other variations based on the technical concept of the present invention are possible in addition to the embodiments disclosed herein.

[0144] [Explanation of the symbol]

[0145] 100, 900: Content creation device

[0146] 110: Trend Collection Department

[0147] 120: Trend Normalization Section

[0148] 130: Trend Evaluation Department

[0149] 140: Instruction generation section

[0150] 150: Content Synthesis Section

[0151] 160: Verification Department

[0152] 170: Feedback Department

[0153] 910: Memory

[0154] 920: Communication interface

[0155] 930: Processor

Claims

Claim 1 A content generation device comprising: a trend normalization unit that converts trend signals collected from multiple sources into a common schema including the meaning and temporal change amount of the trend; a trend evaluation unit that calculates usability for the converted trend based on a lifecycle and risk level based on the temporal change amount; a guideline generation unit that generates content generation guidelines for trends where the usability meets criteria; and a content synthesis unit that generates candidate content based on the content generation guidelines, wherein the content generation guidelines are guidelines that connect the trend with the assets of a target knowledge base. Claim 2 In claim 1, the content generating device further comprises a verification unit that verifies whether the candidate content satisfies a preset condition and classifies the candidate content as content awaiting publication or unpublished content according to the verification result. Claim 3 A content generation device according to paragraph 2, wherein the verification unit calculates a differentiation score based on similarity with reference content for the candidate content, and the preset condition includes the differentiation score being greater than or equal to a standard. Claim 4 A content generation device according to paragraph 3, wherein the differentiation score is calculated based on at least one of string similarity for a title or introductory phrase, n-gram redundancy, and embedding similarity in a representation space between the candidate content and the reference content. Claim 5 A content generation device according to paragraph 3, wherein the verification unit verifies the factual accuracy of the candidate content or whether there is a misunderstanding of the trend context using at least one of the trend source text, news summary, internal knowledge base, and approved source documents, and further uses the verification result as the preset condition. Claim 6 A content generation device according to paragraph 4, wherein the differentiation score is calculated by further considering the degree to which the assets of the target knowledge base are reflected in the candidate content. Claim 7 A content creation device according to claim 1, wherein the trend evaluation unit calculates the usability by combining two or more of the relevance score with the target knowledge base, the rate of increase score, the lifecycle timing score, the sensitivity score, the legal risk score, and the competitive congestion. Claim 8 A content generation device according to claim 1, wherein the trend evaluation unit classifies the lifecycle stage of the trend using two or more of current interest, rising speed, rising acceleration, and observation duration from the hourly observation value of the trend signal. Claim 9 In paragraph 3, the content generating device further comprises a feedback unit that records the performance of the published content and adjusts at least one of the criteria for calculating the usability, the criteria for calculating the differentiation score, and the publishing timing for each lifecycle stage based on the performance. Claim 10 In claim 8, the content synthesis unit further calculates a publication time score of the candidate content based on the lifecycle stage of the trend, a content generation device. Claim 11 A content generation method performed by a content generation device, comprising: a step of converting trend signals collected from multiple sources into a common schema including the meaning and temporal change amount of the trend; a step of calculating usability for the converted trend based on a lifecycle and risk based on the temporal change amount; a step of generating content generation guidelines for trends in which the usability meets criteria; and a step of generating candidate content based on the content generation guidelines, wherein the content generation guidelines are guidelines for connecting the trend with an asset of a target knowledge base. Claim 12 In claim 11, the content generation method further comprises: a step of verifying whether the candidate content satisfies a preset condition; and a step of classifying the candidate content into content awaiting publication or unpublished content according to the verification result. Claim 13 A content generation method according to claim 12, wherein the verification step comprises a step of calculating a differentiation score based on similarity with reference content for the candidate content, and the preset condition comprises that the differentiation score is greater than or equal to a standard. Claim 14 A content generation method according to claim 13, wherein the differentiation score is calculated based on at least one of string similarity for a title or introductory phrase between the candidate content and the reference content, n-gram redundancy, and embedding similarity in a representation space. Claim 15 A content generation method according to claim 13, wherein the verification step comprises verifying the factual accuracy of the candidate content or whether there is a misunderstanding of the trend context using at least one of the trend source text, news summary, internal knowledge base, and approved source documents, and the verification result is further used as the pre-set condition. Claim 16 A content creation method according to claim 14, wherein the differentiation score is calculated by further considering the degree to which the assets of the target knowledge base are reflected in the candidate content. Claim 17 A content creation method according to claim 11, wherein the step of calculating the usability comprises the step of calculating the usability by combining two or more of the relevance score with the target knowledge base, the rate of increase score, the lifecycle timing score, the sensitivity score, the legal risk score, and the competitive congestion. Claim 18 A content creation method according to claim 11, wherein the step of calculating the usability includes the step of classifying the lifecycle stage of the trend using two or more of current interest, rising speed, rising acceleration, and observation duration from the hourly observations of the trend signal. Claim 19 In claim 13, the content creation method further comprises: a step of recording the performance of the posted content; and a step of adjusting at least one of the criteria for calculating the usability, the criteria for calculating the differentiation score, and the posting timing for each lifecycle stage based on the performance. Claim 20 A content generation method according to claim 18, wherein the step of generating the candidate content further includes the step of calculating a publication point score of the candidate content based on the lifecycle stage of the trend.

Citation Information

Patent Citations

  • AI garment style customization design method based on multi-modal user portrait

    CN121350998A

  • Systems and methods for analyzing and evaluating trends in senior content

    KR1020240097181A

  • Method for analyzing technology and device thereof

    KR1020260111227A