Trademark risk management system and method

Abstract

The present invention provides a trademark risk management system and method implemented by providing an electronic device operated by a user, the electronic device including a processor and a network interface controller, a server including an application, the processor connected to the server via the network interface controller, executing the application and used for category recommendation and risk management purposes, in particular for recreating text or graphics.

Description

[Technical field]

[0001] The present invention relates to a trademark risk management system and method, which is particularly used for recreating trademark text or graphics. [Background technology]

[0002] In the traditional patent application process in China and internationally, both paper documents need to be printed and multiple forms need to be filled out. Many documents are paper, which is a huge hassle in terms of management and classification. Not only is it environmentally unfriendly, it also wastes a lot of paper and may lead to errors in the patent application process or invalidation of patents due to supervision, resulting in serious losses. In addition to the traditional paper preparation, if communication between personnel is required in the patent application process, information may be inaccurately transmitted or misunderstood due to their different professional backgrounds, language usage, cultural differences and other unpredictable factors. This leads to understanding differences between applicants, offices, agents and government agencies. As a result, applicants may not be able to achieve the results they originally intended.

[0003] In addition, patent applications are intended to be a symbol of a company's image, as well as a defensive weapon in patent infringement lawsuits. In fact, the value of a patent right can be created not only by allowing others to use the patent to obtain license income, but also by the right to eliminate infringement and the right to compensation for damages, which can generate financial revenue.

[0004] For enterprises, protecting the results of their research and development through patent applications has always been a necessary and important part of the business process. Some enterprises seem to think that patent applications are the job of professional patent offices. Professional patent offices can draft patent documents with accurate technical descriptions, complete content disclosure, and broad scope of inventions by inviting contractors to discuss the technical contents of patent applications.

[0005] In fact, patent filing is not only a collaborative task between an enterprise and a patent office, but also communication between different departments within an enterprise, especially between the R&D department and the IP department, in the repeated communication regarding the enterprise's internal patent filing, the R&D engineers in the R&D department need to provide the relevant technical content. This may include the description of the basic background of the technology, the defects or problems in the prior art that need improvement, and the characteristics of the new technology. In addition, it is necessary to search the proposed technical content in advance in the patent databases of multiple countries to identify similar prior art, and facilitate the discussion of the internal patent proposal. In many enterprises, the number of people in the R&D department far exceeds the number of people in the internal IP department, which increases the workload of the IP department.

[0006] However, although many medium and large enterprises have a proper related patent proposal system, in reality, due to the professional differences between the development department and the intellectual property department, there is a lot of time cost, from several days to several months, in communication. Although the development department is familiar with the technology, it usually cannot meet the requirements of the intellectual property department required for patent proposal (patent disclosure). In contrast, when discussing with the development department based on the prior technology search report, the intellectual property department cannot explain the differences clearly and in detail for the development department to understand at a glance. Therefore, the progress of internal patent proposal (patent disclosure) discussion requires a lot of effort and time. If a patent disclosure including drawings and even patent claims is prepared within the company, it usually requires at least several days to several weeks. Many companies outsource to third-party patent and trademark offices or law firms to cooperate in preparing patent specifications and filing patent applications without patent disclosure, which incurs more costs in terms of communication and understanding, such as developers communicating about technology again, and the time for patent application is delayed, affecting the technological rights of the company.

[0007] Also, for small and medium-sized enterprises and start-ups without intellectual property departments, patent searches are entirely left to third-party patent and trademark offices and law firms. However, since there are no patent proposals (patent disclosures), this process is mainly carried out by presentations or chats. Such a model generally causes inventors to spend a lot of effort and time on technical communication. The communication and understanding costs in this process are astonishing compared to medium and large enterprises.

[0008] However, many search technologies and platforms available today help reduce user interaction time.

[0009] The patent drafting support system disclosed in Chinese Application No. CN201610297330.0 includes: a disclosure template creation module for creating a template for drafting a technical disclosure document including a plurality of fields; a disclosure input module for inputting content corresponding to the plurality of fields of the template to create a technical disclosure document; a content identification and extraction module for identifying and extracting content from each field of the technical disclosure document; a related search module connected to an external database server, for selecting content extracted by the content identification and extraction module and associating it with the external database server so as to search for related information from an external database; a copy storage module for copying and storing data searched from the external database; and a document creation module for creating a document in a predetermined format using the copied data.

[0010] However, the above Chinese application number CN201610297330.0 has some problems to be improved. Its main features are to modularize the text entered by the author into corresponding fields, extract keywords from the modularized text based on different fields, search each keyword, and finally create search data for the author to refer to. This reduces the time of manual keyword search input. The actual time reduction for document drafting, analysis and comparison, and communication between different departments in the enterprise is quite limited, and it does not support graph generation and is only applied to text.

[0011] For example, as disclosed in Taiwan Application No. TW097119308, there is a system for preparing a patent specification, which includes a method mainly implemented as software or an application. The present invention relates to a system capable of preparing a written paper copy of a patent specification. The above-mentioned system includes: a central processing means for executing a calculation process of the method for preparing the patent specification; a data storage means connected to the central processing means and storing the method for preparing the patent specification; an input means connected to the central processing means and providing a user with an interface for inputting related art disclosure data; a control means connected to the central processing means and controlling the process of preparing the data of the patent specification; an output means connected to the central processing means and used as an interface for outputting the data of the patent specification, which can be connected to a printing means for printing the patent specification; and finally a display means connected to the central processing means and displaying the outputted data of the patent specification.

[0012] However, the above Taiwan Application No. TW097119308 has some problems to be improved. The relevant text needs to be input into the relevant field. This method is mainly related to the use of general terms in patent specifications, and is similar to filling in blanks to create the body of the specification. However, the input text must conform to the pre-set specifications, otherwise the sentences created may be grammatically inaccurate, making it difficult for general users to start using it easily. In essence, users have to put in a lot of time to create documents themselves, and the problem of correspondence documents disclosed in patents before filing cannot be solved. It is also limited to text and cannot support the creation of figures.

[0013] As disclosed in U.S. Patent Application No. US17745671, in certain embodiments, the assistance system may assist a user in obtaining information or services. The assistance system allows a user to interact in a stateful multi-round session with various input modes (e.g., audio, voice, text, graphics, video, gestures, movements, position, orientation, etc.) to receive assistance from the system. By way of example and not limitation, the system may support single-modal input (e.g., voice only), multi-modal input (e.g., voice and text), hybrid / multi-modal input, or any combination thereof. The user input provided may be associated with a particular assistance task and may include, for example, a user request (e.g., a verbal request for information or performing a movement), an interaction with an assistance application associated with the assistance system (e.g., selection of a UI element by touch or gesture), or any other type of suitable user input that can be detected and understood by the assistance system (e.g., a user movement detected by the user's client device). The assistance system may create and store a user profile that includes personal and contextual information associated with the user. In some embodiments, the assistance system may parse the user input using natural language understanding (NLU). This analysis may provide a more personalized and context-aware understanding based on the user's profile. The system may analyze entities associated with the user input based on the analysis. In some embodiments, the assistance system may interact with different agents to obtain information or services associated with the analyzed entities. The assistance system may generate responses for the user regarding information or services using natural language generation (NLG).

[0014] However, the above US Patent Application No. US17745671 has some problems to be improved. In the previous situation, natural language understanding and natural language creation are used to analyze the user input text and create the corresponding reaction action or response realized by the "voice assistant". However, the language used or trained in the background data, which is only applied to general daily life, is quite different from the legal language of intellectual property (patent and trademark) in terms and grammar. When the horizontal training mode is applied to the vertical industry of intellectual property, the accuracy is significantly reduced.

[0015] For example, as disclosed in the doctoral dissertation of Li Jiesheng of the Institute of Computer Science and Engineering, National Taiwan University (Application of Deep Learning in Patent Field 10.6342 / NTU202100999), there are problems of insufficient training scale of language model and insufficient data amount, which may lead to inaccurate training results. In addition, the training is not deeply focused on the patent language. This method relies heavily on the accuracy of the input text to create a summary and derive the technical content from the summary. This makes it easy for the text of the technical content to be biased. The comparison method used in this paper is text mapping, which has many advantages, such as capturing the semantic information and contextual relationship of the text, but has some disadvantages as follows: (1) Dimensional disaster: For very large text data, text mapping may cause high-dimensional vector representation, which is challenging to calculate and store. (2) Semantic identification is difficult: Text mapping may capture some semantic information, but may not fully reflect the meaning of ambiguous or context-related words. Semantic similarity between different texts may also be a problem. (3) Training Data Requirement: Text mapping methods typically require a large amount of labeled training data to learn vector representations of words and text. Obtaining large-scale, high-quality labeled data can be time-consuming and expensive. (4) Vocabulary Maintenance: Although text mapping methods map words to vector spaces, they may encounter new vocabulary, typos, or other word variants in real applications. To ensure the accuracy of the model, vocabulary maintenance and updating may be necessary. In addition, training is performed using BERT and version 2.0 in this specification, which has limitations. The communication problem between inventors and patent employees has not been resolved.

[0016] In addition, current natural language processing (NLP) models such as Chat GPT (registered trademark) 3.5 and 4.0 developed by OpenAI are autoregressive language models trained by reinforcement learning with human feedback (RLHF). They are mainly used for customer service conversation processing, story explanation, translation, grammar editing, poetry, song lyrics, text organization, and even software program creation. However, when an inventor uses GPT to create a patent disclosure based on a dialogue, it cannot directly create accurate patent industry content or create a related patent process diagram. For inventors, it is only used as a chatbot to consult patent regulations. In addition, when a search on patents is performed using GPT, only the correct patent certificate number is created, and accurate case names and patent charts are not provided. See Figures 1 and 2. For patent employees, it is necessary to repeatedly verify, which may be time-consuming. Therefore, a deep learning model that focuses on the intellectual property vertical and aims for industry accuracy is needed to help manage patent disclosures and patent searches and solve this problem.

[0017] For example, like traditional patent drafting tools such as Patent Theory, PowerPatent, Rowan, etc., these tools help to prepare patent documents, and provide necessary assistance or editing functions with the assistance of professionals during the drafting process. They also contribute to maintaining consistency in terminology, such as reducing the error rate in the completed patent description, and thus to creating written suggestions or contents for reference. Meanwhile, the efficiency of preparing patent specifications has also been improved.

[0018] However, the above conventional tools still have some problems that need to be improved. These tools are mainly developed for professionals in the intellectual property industry. That is, the main target users of these tools are engineers who have experience in writing patent specifications. These tools only play a supporting role. Therefore, relatively inexperienced developers in companies can hardly use these tools to help them solve problems. The operation of these tools is not intuitive for them, and these tools cannot create graphs. These tools cannot create technical disclosures or graphs for non-professional individuals or personnel from small and medium-sized enterprises.

[0019] From the above description, it is clear that the known technology needs to be improved or adjusted in order to improve the convenience of developers in choosing trademarks. In order to avoid missing the timely opportunity for urgent trademark applications, the inventors have made a great deal of consideration and ingenuity, and finally developed the system and method according to the present invention. Summary of the Invention [Problem to be solved by the invention]

[0020] The present invention aims to provide a system and method of implementation that solves problems in the prior art. [Means for solving the problem]

[0021] Therefore, to achieve the object of the present invention, the present invention provides a trademark risk management method realized by providing an electronic device operated by a user, the electronic device includes a processor and a network interface controller. There is a server including an application. The processor is connected to the server through the network interface controller and executes the application to perform category recommendation and risk management, and the method includes at least:

[0022] (S100) a user causes a processor to execute an input module to input text content via a user interface of an electronic device;

[0023] (S200) the processor executes a semantic analysis module in the application to perform semantic analysis on the text content;

[0024] (S300) the semantic analysis module is further connected to the classification module, the search module and the database module, and classifies the analyzed text content according to industry technology, performs a comparison search in the database module, and sends the matching data to the intellectual property information disclosure module;

[0025] (S400) The intellectual property information disclosure module analyzes and compiles the data and presents the information to the user via the user interface of the electronic device.

[0026] (S500) The analyzed and compiled data is also sent to a category recommendation module, which classifies and compiles the intellectual property in the data, and displays the application intellectual property types to be ranked and recommended on a user interface.

[0027] (S600) a user selects a trademark type via a user interface;

[0028] (S700) a user inputs a brand description via a user interface and presents it to a processor to execute an input module;

[0029] (S800) The login module completes the login, and the processor executes the semantic analysis module and the search module to match the analysis result in the database module and generate a recommended trademark category;

[0030] (S900) The processor includes a step of executing the search module to perform a second comparative search in the database module based on the recommended trademark categories and to send the second search results to a search document creation module that creates a risk assessment report.

[0031] After step (S900), the process further includes a step (S901) in which the risk management module recreates the text or figure based on similar text or figure in the risk assessment report and the concept information from the user received by the input module.

[0032] It is another object of the present invention to provide a system that overcomes the problems in the prior art.

[0033] Therefore, in order to achieve another object of the present invention, there is provided a trademark risk management system operated by a user through an electronic device. The processor of the device is connected to a server through a network interface controller and executes an application for category recommendation and risk management. The system includes at least:

[0034] an input module for receiving and marking text content input by a user, converting the text content into a character string, transmitting the character string information, and recording the input language of the character string information in a temporary memory;

[0035] a semantic analysis module that receives the character string information, analyzes it using a natural language database, and marks it, and creates and transmits a semantic analysis result;

[0036] a classification module that analyzes the results of the semantic analysis of the industry category code, and is connected to the database module to determine and create at least one set of industry classification codes;

[0037] a search module that performs a comparative search in the database module based on at least one set of industry classification codes and generates search result data;

[0038] an intellectual property information disclosure module that receives the data, performs further statistical analysis, and generates basic intellectual property information;

[0039] a recommendation module that receives the data and classifies and compiles the intellectual property rights in the data to create and apply recommended types of intellectual property rights;

[0040] and a login module for the user to authenticate himself / herself on the electronic device.

[0041] The input module receives the trademark description input by the user for the second time when the trademark is selected from the intellectual property type recommended by the user. The semantic analysis module receives and analyzes the string information related to the brand description after the identity is verified by the login module, and transmits the analysis result of the technical description. The search module performs a comparative search in the database module using the analysis result, and transmits the search result to the recommendation module to create a recommended trademark category. The search module performs a second comparative search in the database module based on the recommended trademark category, and transmits the second search result to the search document creation module to create a risk assessment report.

[0042] It further includes a risk management module that recreates the text or graphics based on similar text or graphics in the risk assessment report and the concept information received from the user by the input module.

[0043] The following detailed description is provided by way of specific examples only and is further illustrated by the drawings. [Brief description of the drawings]

[0044] [Figure 1] FIG. 1 is a schematic diagram of the prior art. [Diagram 2] FIG. 2 is another schematic diagram of the prior art. [Diagram 3]1 shows a flow chart of a method according to the present invention. [Figure 4] 1 is a schematic diagram of a system according to the present invention; [Diagram 5] 1 is a schematic diagram of a system according to the present invention; [Figure 6] 1 is a schematic diagram of a system according to the present invention; [Figure 7] 2 shows a schematic diagram of another embodiment of a system according to the present invention; [Figure 8] 1 shows a flow chart of a method according to the present invention. [Figure 9] 1 shows a flow chart of a method according to the present invention. [Figure 10] 1 shows a flow chart of a method according to the present invention. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0045] The present invention provides a trademark risk management system and method. The method according to the present invention is executed by a user operating an electronic device, as shown in Fig. 3. The processor of the electronic device is connected to a server via a network interface controller and executes an application for category recommendation and risk management. The method includes at least:

[0046] (S100) a user instructs a processor to execute an input module to input text content via a user interface of an electronic device;

[0047] (S200) the processor executes a semantic analysis module of the application to perform semantic analysis on the text content;

[0048] (S200) the processor executes a semantic analysis module of the application to perform semantic analysis on the text content;

[0049] (S300) the semantic analysis module is further connected to the classification module, the search module and the database module, for classifying the analyzed text content as industrial technology, searching in the database, and sending the obtained data to the intellectual property information disclosure module;

[0050] (S400) the intellectual property information disclosure module performs statistical analysis on the data and presents the information to the user via a user interface of the electronic device;

[0051] (S500) the analyzed and statistically processed data is also sent to a category recommendation module, which classifies and statistically analyzes the intellectual property types in the data, and displays and applies the recommended intellectual property types on a user interface;

[0052] (S600) a user selects a trademark type via a user interface;

[0053] (S700) a user instructs a processor to execute an input module to input a brand description using a user interface;

[0054] (S800) a user completes login through a login module, a processor executes a semantic analysis module, the search module searches in the database based on the analysis result, and the search result is sent to the recommendation module to generate a recommended trademark category;

[0055] (S900) The processor includes a step of executing the search module to perform a second search in the database based on the recommended trademark category, and sending the second search results to a document search module to generate a risk assessment report.

[0056] After step (S900), the risk management module further includes a step of (S901) recreating text or graphics based on similar text or graphics in the risk assessment report and concept information from the user received by the input module.

[0057] For an illustration of a system according to the present invention, please refer to FIGS.

[0058] The system according to the present invention is realized by providing an electronic device 100 operated by a user. The electronic device 100 includes a processor 101 and a network interface controller 102. The server 200 includes an application 201. The processor 101 is connected to the server 200 via the network interface controller 102 and executes the application 201 for category recommendation and risk management. The system includes at least a login module 300, an order processing module 301, an input module 302, a semantic analysis module 303, a classification module 304, a content learning module 305, a risk management module 306, a diagram creation means 309, a text creation means 3091, a document search module 310, a diagram search means 311, a text search means 3111, a database module 600, a natural language database 500, a recommendation module 802, a category recommendation module 700, and an intellectual property information disclosure module 703.

[0059] When the user selects trademark as the recommended intellectual property type in the recommendation application, the input module then receives the user's input of the brand description again and completes identity verification through the login module.Then, the semantic analysis module receives the string information about the brand description, performs analysis and word segmentation to create and send the analysis result of the technical description.Then, the search module compares the analysis results in the database module and sends the search result to the recommendation module to create a trademark category for the recommendation application.The search module performs another comparative search in the database module based on the trademark category of the recommendation application, and sends the second search result to the search document creation module to create a risk assessment report.

[0060] The risk management module recreates the text or graphic based on similar text or graphic from the risk assessment report and the concept information received from the user by the input module.

[0061] Specifically, a user may use an electronic device 100 to connect to the server 200 via the network interface controller 102 and execute an application 201. They may input text or a website link. The input text may be a description of a corporate brand or a technical overview of a core product or service of the corporate brand. They may input a URL link of an official website. The semantic analysis module 303 may perform semantic analysis on the content of the website and the input text by extracting and reading the content of the website, and send the analysis result to the classification module 304.

[0062] The classification module 304 performs industry category code analysis on the semantic analysis result, and is connected to the database module to determine and generate at least one set of trademark classification codes.

[0063] The classification module then connects to the database module 600 using at least one set of patent classification codes to perform a comparative analysis, such as comparing information such as the number of intellectual property applications in the database module 600 and categories related to the same or similar industries. This comparison generates data related to search results.

[0064] The intellectual property information disclosure module receives the data and performs further statistical analysis to generate basic intellectual property information. This disclosed basic information may include details such as technology value, technology risk, intellectual property time cost, and payment cost.

[0065] The present invention provides a trademark risk management system and method that can automatically recreate trademark names and designs based on the inventor's creative or conceptual input. The system of the present invention can use a natural language processing (NLP) algorithm to perform an artificial intelligence semantic search when an inventor presents or inputs a creative idea, and convert the inventive step of the input into a programming language for recreating the design, while simultaneously searching and comparing the designs.

[0066] In one embodiment of the present invention, the main point is to describe the process by which a user completes intelligent recommendation and continues with category recommendation, risk assessment, and risk management. However, in this embodiment, the user may also start by receiving a recommendation for an intellectual property type before continuing with category recommendation and risk management. For a complete understanding, please refer to Figs. 4 to 6. The system according to the present invention is realized by an electronic device 100 including a processor 101 and a network interface controller 102. A server 200 includes an application 201. The processor 101 is connected to the server 200 via the network interface controller 102 and executes the application 201 for category recommendation and risk management. The system includes at least a login module 300, an order processing module 301, an input module 302, a semantic analysis module 303, a classification module 304, a category recommendation module 700, a content learning module 305, a graphic learning means 3051, a risk management module 306, a text creation means 3091, a graphic creation means 309, a document search module 310, a text search means 3111, a graphic search means 311, a temporary memory 400, a database module 600, and a natural language database 500.

[0067] The server 200 may be a cloud server or a locally managed server architecture.

[0068] In this embodiment, the creator uses a server with the following specifications to train and execute the trademark item conversion model of this technology: For the processor 101 (CPU), it is a high-performance multi-core processor with at least 16 or more cores, such as AMD Ryzen® Threadripper® or Intel Xeon® series, and such a selection of CPU is particularly important for processing large amounts of data and performing complex calculations. For the memory (RAM), the memory size may be 64 GB or more to fit the size of the corpus data and the word vector model. Network interface controller 102: As a fast and reliable network connection hardware, it is used especially with cloud computing resources or when transmitting large amounts of data. For the GPU (Graphics Processing Unit), a highly efficient GPU such as NVIDIA's RTX30 series or Tesla® series is used to reduce the training time of the model. In such a case, the training server used in the architecture of the server 200 uses the above higher specifications for training purposes. However, once the model training is completed and inference processing is performed, a server host with lower specifications may be used, whether in the cloud or locally. The choice of server host specification does not affect the technical features emphasized in this case, and any server host specification should fall within this technical scope.

[0069] The login module 300 allows a user to authenticate themselves when operating the electronic device 100. It verifies the user's identity while verifying their identification information.

[0070] The order processing module 301 provides users with functions such as member registration / login, creation and management of orders for international patent cases, input of international patent cases, etc. It also provides functions for case order inquiries, confirmation of patent and trademark examination status, and account inquiries.

[0071] A user may register and log in as a member of the system to specifically manage patent applications and case orders on the platform, including tasks such as document creation, case receipt, case copy submission, and formal document notifications.

[0072] A user may create a new case order in the order processing module 301. Once the user has selected a first target country, the system's temporary memory 400 is periodically updated with information from the case order.

[0073] The input module 302 is responsible for receiving explanatory text or graphics input by the user, especially their creative concepts, converting the explanatory text into character strings for marking and transmitting the character string information, and recording the input language in the temporary memory 400.

[0074] Specific implementations may include text preprocessing, tag extraction, serialization, and tag processing. Text preprocessing further includes tokenization, stop word removal, part of speech tagging, and part of speech word tagging. Tag extraction first uses predefined rules and patterns to extract tags, and then uses a machine learning model to perform automatic tag extraction. Columnization may be performed by consolidating the extracted labels into a single column, such as "Movie#Action#Drama". Tag processing may be performed based on specific requirements, and may include removing duplicate tags, classifying tags, and other operations. More specific methods may use natural language processing (NLP) toolkits such as NLTK, Stanford NLP, HanLP, machine learning frameworks, such as TensorFlow, PyTorch, and scikit-learn.

[0075] The semantic analysis module 303 receives character string information, and handles it by performing analysis and marking via the natural language database 500. Then, it creates and transmits the results of the semantic analysis.

[0076] The classification module 304 analyzes the trademark category code based on the result of the semantic analysis and is connected to the database module 600 to determine and create at least one set of trademark classification codes or trademark categories. The classification analysis may include the internationally recognized Nice Classification of Trademarks and / or trademark classifications specific to various countries, such as the Taiwan Trademark Classification.

[0077] To perform classification analysis to generate trademark categories for a paragraph of descriptive text, the following method may be considered:

[0078] Text Preprocessing: First, the input description text is preprocessed, which may involve removing punctuation, stop words and meaningless characters, as well as performing lexical normalization (e.g. stemming or semanticization).

[0079] Feature Extraction: Extracting meaningful features from the preprocessed text. This can be achieved using common text feature representation methods such as bag-of-words or word embeddings. Bag-of-words represents text as word frequency or occurrence, while word embedding maps words into a continuous vector space.

[0080] Trademark Classification Model: Build an industry classification model, which may be based on machine learning classifiers such as support vector machines, decision trees, deep learning models, or rule-based methods, and train the model to classify text into corresponding industry categories.

[0081] Trademark Category Creation: A pre-trained classification model is used to classify the input description text and generate a corresponding trademark classification code. The Nice Classification is an international standard used for classifying trademarks.

[0082] It is assumed to have explanatory text.

[0083] The company specializes in the development of new materials and technologies for photovoltaic power generation. Highly efficient solar cells are designed to convert solar energy into electrical energy while reducing energy costs and environmental impact. These technologies have potential in the renewable energy sector and are widely applied for residential and industrial purposes.

[0084] A classification analysis of a trademark may be performed to create trademark categories by the following steps:

[0085] Text preprocessing: Preprocess the explanatory text by removing punctuation, converting to lower case, removing stop words, etc. For example, the above explanatory text is preprocessed as follows: "Company specializes in developing new materials and new technologies for photovoltaic power generation High-efficiency solar cell design Solar energy conversion Energy cost reduction Environmental impact reduction Renewable energy field Potential Widely applied for residential and industrial purposes."

[0086] Feature Extraction: The preprocessed text is converted into a feature vector representation using a Bag-of-words model. Each word may be considered as a feature, and the frequency of a word in the explanatory text may be used to represent its importance. For example, the following feature vector representation can be obtained:

[0087] {"Company":1, "Specialization":1, "Research":1, "Development":1, "Solar":2, "Power Generation":1, "New":1, "Material":1, "Technology":2, "Design":1, "High Efficiency":1, "Battery":1, "Conversion":1, "Power":1, "Reduction":1, "Energy":1, "Cost":1, "Environment":1, "Impact":1, "Renewable Energy":1, "Potential":1, "Widely":1, "Application":1, "Home":1, "Industry":1, "Application":1}.

[0088] Trademark Classification Model: Using the marked training data with corresponding labels for model training, build a trained classification model, such as a support vector machine or a deep learning model, such as a recurrent neural network or a convolutional neural network.

[0089] Category Creation: The trained classification model is used to input the pre-processed text feature vector and make a prediction. Based on the prediction result, a corresponding trademark category may be created. For example, the model may predict that the text of the specification belongs to the solar energy related industry category, and the corresponding Nice classifications are categories 04, 09, 37, and 40, etc.

[0090] The category recommendation module 700 uses a computation model trained by a natural language model, a trademark classification table and sub-categories, and parses string information with ambiguous meaning (or imprecise description) into trademark category recommendation information, and finally combines with the semantic analysis module 303 and the classification module 304 to generate recommended trademark categories for applications.

[0091] Natural language models can predict the next word or create sentences that conform to grammar and meaning based on known text data, including but not limited to:

[0092] N-gram model: N-gram model is a probabilistic language model that assumes that the probability of a word occurring depends only on the previous N-1 words. For example, a binary model (N=2) predicts the probability of the next word based on the previous words.

[0093] Recurrent Neural Network (RNN) Model: RNNs are neural networks well suited to dealing with sequence data, allowing them to capture the time dependencies between words. In natural language processing, RNNs are typically used to build language models that consider each word as a time step.

[0094] Pre-trained language models (e.g., BERT): A pre-trained language model is a model that is trained on a large, unsupervised dataset to understand and create natural language. BERT models, based on a converter architecture, are pre-trained on a wide range of text data and then fine-tuned for a specific task (e.g., text classification, named entity recognition, etc.).

[0095] Trademark classification tables and subcategories are extracted from national trademark databases, including all goods and service categories listed in these classifications, which are extracted and compiled into a list that is then used for training and computation by the natural language model.

[0096] The document search module 310 includes a text search means 3111, a graphic search means 311, a conversion means 317, and a graphic matching means 318. The graphic search means 311 receives a graphic input by a user and searches in the database module 600, and the text search means 3111 receives an input text and searches in the same database module 600 to obtain a searched document.

[0097] Specifically, when the graphic search unit 311 receives an input graphic, it first converts the graphic using the conversion unit 317 and adds Wien classification tags. At the same time, the graphic is converted into a vector. Then, the graphic matching unit 318 compares it with the database module 600 and creates a search document.

[0098] Convert the shape into a vector, and assume for example that the shape is basically made up of pixels, with the brightest pixels considered to be 1 and the darkest pixels considered to be 0, so don't take the primary colors into account. For a 3x3 pixel shape, top left, center left, top center, center left, center middle, center right... bottom right corner is (1, 0, 1, 0, 1, 0, 1, 0, 1). In this case the shape is black-white-black-white-black-white-black-white-black-white-black. Taking into account the three primary colors, we can simply extend the length of this vector by a factor of three and transform it to R(1,0,1,0,1,0,1,0,1)+G(1,0,1,0,1,0,1)+B(1,0,1,0,1,0,1,0,1,0,1)=(1,0, ...

[0099] Specifically, during the matching process, similarity measures may be calculated in different forms, such as edit distance, cosine similarity, Jaccard similarity, etc. The figures having a similarity level above a set threshold may be filtered out, and this threshold may be adjusted as necessary.

[0100] Edit distance is typically used to measure similarity between strings, quantifying the minimum number of operands (insertions, deletions, substitutions) required to transform one string into another, but it can also be used to measure similarity between approximate vectors.

[0101] Below are the general steps to calculate vector similarity using edit distance.

[0102] Vector serialization: Convert each vector into a sequence. For example, a digital vector may be directly converted to a digital sequence. A feature vector may combine the name and value of each feature into a sequence of columns.

[0103] Calculate edit distance: Calculate the edit distance between the two sequences using an edit distance algorithm (e.g., Levenshtein distance).

[0104] Normalization: For ease of comparison, we normalize the edit distances to have values ​​ranging from 0 to 1. A common normalization method is to divide by the maximum length of the sequence.

[0105] Similarity: indicates the degree of similarity between two vectors (1-normalized edit distance). Typically, the closer the value is to 1, the more similar the vectors are.

[0106] For example, assume there are two vectors: Vector 1 = [1, 2, 3, 4] and Vector 2 = [2, 3, 1, 4].

[0107] Serialize: Vector 1 -> "1234", Vector 2 -> "2314"

[0108] Edit distance: 2 (requires two operations to swap "1" and "3")

[0109] Normalization: 2 / 4=0.5

[0110] Similarity: 1-0.5=0.5

[0111] Therefore, the similarity between the two vectors is 0.5, indicating a certain degree of similarity.

[0112] Cosine similarity is a method for measuring the directional similarity between vectors, and is used to estimate the similarity between vectors. The principle is to calculate the cosine value of the angle between two vectors. The closer the angle is to 0, the more similar the vectors are, and the closer the cosine value is to 1.

[0113] Below are general steps to calculate vector similarity using cosine similarity.

[0114] Vector preprocessing: ensure that two vectors have the same dimensions and normalize the vectors (e.g., make their length 1) to remove effects due to vector length differences.

[0115] Dot product calculation: Calculate the dot product between two vectors, multiply corresponding elements pairwise, then sum them. The dot product reflects the directional correlation between the two vectors.

[0116] Calculating vector length: This involves calculating the length of each vector, summing the squares of the corresponding elements, and then taking the square root. The length of a vector reflects its overall magnitude.

[0117] Calculating cosine similarity: To get the cosine similarity between two vectors, divide the dot product by the product of the lengths of the two vectors.

[0118] Similarity Analysis: Cosine similarity has a value range of -1 to 1. When the cosine similarity is close to 1, it indicates that the two vectors have very similar directions and are highly similar. When the cosine similarity is 0, it means that the two vectors are orthogonal and not similar at all. When the cosine similarity is close to -1, it indicates that the two vectors have opposite directions and are not similar in height.

[0119] Assume there are two vectors: Vector 1 = [2, 3, 4] and Vector 2 = [4, 6, 8].

[0120] Vector Preprocessing: Since these two vectors have the same dimensions, no further processing is required.

[0121] Calculating the dot product: 24+36+4*8=62

[0122] Calculating vector length: sqrt(2^2+3^2+4^2)=sqrt(29), sqrt(4^2+6^2+8^2)=sqrt(100).

[0123] Calculating cosine similarity: 62 / (sqrt(29)*sqrt(100))≒0.906

[0124] Thus, the cosine similarity between the two vectors is close to 0.9, indicating a very high level of similarity in direction.

[0125] Jaccard similarity is a method to measure the similarity between vectors and is used to estimate the similarity between vectors. The principle is to calculate the ratio of common elements between two vectors. A high ratio indicates a high similarity, resulting in a high Jaccard similarity.

[0126] Below are the general steps to calculate vector similarity using Jaccard similarity.

[0127] Vector preprocessing: Ensure that two vectors have the same dimensions. Normalize the vectors (e.g., make their elements sum to 1) to remove effects of vector length differences.

[0128] Intersection Counting the number of common elements between two vectors.

[0129] Finding a union: Count the total number of elements in two vectors.

[0130] Calculating Jaccard similarity: To get the Jaccard similarity between two vectors, divide the intersection by the union.

[0131] Similarity Analysis: Jaccard similarity has a value range from 0 to 1. A Jaccard similarity close to 1 indicates that the two vectors are the same and highly similar. A Jaccard similarity of 0 means that the two vectors are completely dissimilar.

[0132] Assume there are two vectors: vector1 = [1, 2, 3, 4] and vector2 = [1, 3, 4, 5].

[0133] Vector Preprocessing: These two vectors have the same dimensions, so no further processing is required.

[0134] Calculating the intersection: 1+3+4=8

[0135] Calculating the union: 1+2+3+4+5=15

[0136] Jaccard similarity calculation: 8 / 15 ≒ 0.533.

[0137] Therefore, the Jaccard similarity between the two vectors is close to 0.5, indicating a certain degree of similarity.

[0138] The text retrieval means 3111 receives the semantic analysis results from the semantic analysis module and performs matching analysis, such as similarity matching analysis of trademark names, in the database module 600 .

[0139] The search document creation module 310 combines the matching results from the graphic search module 311 and the text search module 3111 to create a search document and further classify and filter cases whose similarity is higher than a predefined risk threshold, and generate a risk assessment report.

[0140] The content learning module 305 is a large-scale language model that learns different trademark contents from the database module 600 for different trademark categories. The figure learning means 3051 further learns the trademark figure designs used in different trademark categories, such as figure designs, text designs, unique words, etc.

[0141] The risk management module 306 creates text and / or graphics based on the learning from the content learning module 305 and the past cases matched by the graphics searcher 311 and the text searcher 3111 .

[0142] The database module 600 provides multiple databases required for searching trademark pre-files of multiple countries in the system, such as official trademark databases from each country, trademark category databases from multiple countries, Nice category databases, etc.

[0143] The graphic creation means 309 receives user inputs regarding trademark name or graphic changes and concepts through the input module 302. The graphic creation means combines with the semantic analysis module to analyze the concept and translate it into a graphic creation language. For the concept, a corresponding graphic code is created and then compiled to create a graphic that matches the concept. In the process of recreating a graphic, the graphic search means 311 compares the similarity with past cases retrieved in the past, and ensures that the similarity between the regenerated graphic and the past cases is less than a predetermined threshold.

[0144] The text creation means 3091 receives user input regarding a change or concept of a trademark name or figure through the input module 302. The text creation means 3091 combines with a semantic analysis module to analyze the concept, and combines with a content learning module to create a text which is the trademark name. At the same time, the text search means 3111 compares the similarity between the recreated text and past cases retrieved in the past so that the similarity between the recreated text and past cases is less than a predetermined threshold.

[0145] The risk management module 306 combines the reconstructed concepts to reduce the application risk. After the risk assessment report is generated, the user may choose whether to participate in the risk management by the system. In other words, the user may decide whether to reconstruct text or graphics by the system to avoid text or graphics in past cases.

[0146] Specifically, translating conceptual ideas into graphical language involves several steps. First, the translation module needs to analyze the extracted conceptual ideas to understand the meaning. Then, the translation module needs to graphically generate language generation instructions based on the grammar rules of the language. More specifically, the following steps are used to generate graphical generation language, including technical idea identification, technical feature identification, and technical relationship identification.

[0147] Technical idea identification: First, the translation module identifies technical ideas in the technical information. Technical ideas are basic elements in the diagramming language.

[0148] Technical feature identification: Next, the translation module identifies the technical features in the technical information. The technical features are used to explain the technical idea.

[0149] Technical Relationship Identification: Finally, the translation module identifies technical relationships in the technical information. Technical relationships describe the connections between technical ideas.

[0150] In practice, the semantic analysis module 303 may perform keyword extraction, which is a natural language processing technique that aims to automatically extract important keywords or phrases from text. Methods for extracting keywords may include, but are not limited to, statistical methods, frequency-based methods, text vectorization methods, or machine learning methods.

[0151] Frequency-based methods determine the importance of a word based on its frequency in a text. Common methods include Term Frequency-Inverse Document Frequency (TF-IDF) and Term Frequency. TF-IDF considers the frequency of a word in a text and its importance in the entire document set, while Term Frequency only considers the frequency of a word in a text.

[0152] Text statistical methods rely on statistical models to analyze the distribution and relevance of words in text. Common methods include Mutual Information, Pointwise Mutual Information, and Chi-square test. These methods usually require establishing a statistical model between words and text, and calculating the importance of words based on the model.

[0153] Text vectorization methods convert text into a vector representation and then use a vector space model to calculate the importance of words. Common methods include the Bag-of-words model, word embeddings, and text vectorization methods, such as TF-IDF vectorization.

[0154] Keyword expansion is another feature of keyword extraction that takes the generated keywords, finds similar or synonymous words to expand them, and connects to a synonym library to create additional synonyms.

[0155] The system further includes a language detection module 320 for determining whether the language of the string input by the user is consistent with the official language of the first target country. If the language detection module 320 determines that the string information is not consistent with the official language of the first target country, the translation module 321 translates the string information. After generating the search document, the translation module 321 translates the language of the search document back to the language of the original string information.

[0156] The system also includes a case processing module 312 that executes the system's case orders and prepares conforming trademark application documents. The system further includes a multi-country conversion module 324 that can perform multi-country conversion of trademark application documents and can also integrate the system's case orders with third-party electronic payment services.

[0157] The case processing module 312 further includes an application form creation means 323. The application form creation means 323 extracts and integrates user login authentication ID information into the application data, or the user may directly input the application data received by the input module 302 into the fields by the electronic device 100. Then, the application data is created as an application form according to an application format template.

[0158] After the application is prepared, the user can operate the electronic device 100 to submit the application online or manage the case online through the case processing module 312. The submission is completed by the user directly submitting the application to the authority or by selecting the submission function, which is then handled by the law firm of the providing system, especially taking into account different application document formats and file formats in different countries.

[0159] The system further includes a multi-country conversion module 324. After the application is created, the user may select a second target country through the multi-country conversion module 324 to apply for a trademark right in a different country.

[0160] When the user selects a second target country, the language detection module 320 first determines whether the language of the string information matches the official language of the second target country. If not, the translation module 321 translates the trademark name into the official language of the second target country. Then, after translating the application form into the official language of the second target country, it performs formatting adjustment.

[0161] Alternatively, the language detection module 320 may first determine whether the language of the application form matches the official language of the second target country. If so, it directly adjusts the instruction file and the application form according to the official file format of the second target country, and if not, the translation module 321 translates the application form into the official language of the second target country and then performs formatting adjustments.

[0162] In another embodiment of the present invention, please refer to Figure 4 and Figure 7. Figure 7 shows another schematic diagram of the system according to the present invention, which includes a data receiving module 330, a semantic analysis module 303, a keyword extraction module 331, a search module 332, a classification analysis module 333, a text creation module 334, a content learning module 305, a keyword expansion module 335, a document classification module 336, an automatic writing module 337, and a search report creation module 338.

[0163] The data receiving module 330 may include a login module 300, an order processing module 301, and an input module 302. The login module 300 allows the user to perform identity verification by the electronic device 100, and performs identity verification of the user (login user) and confirms the identification information. The order processing module 301 can provide the user with functions such as membership registration / login, multinational trademark case order creation / management, and multinational trademark case introduction, as well as functions such as case order inquiry, trademark examination status confirmation, and account inquiry. The input module 302 receives explanatory text or figures input by the user, converts the explanatory text into a character string for marking processing, transmits the character string information, and records the input language of the character string information in the temporary memory 400.

[0164] The semantic analysis module 303 receives character string information, analyzes it using the natural language database 500, converts it into markings, and creates and transmits the results of the semantic analysis.

[0165] The keyword extraction module 331 performs word division and keyword extraction based on the input contents, and creates a plurality of keywords.

[0166] Specifically, keyword extraction is a natural language processing technique for automatically extracting important keywords or phrases from text. Methods may include, but are not limited to, statistical methods, frequency-based methods, statistical methods, text vectorization methods, or machine learning methods.

[0167] Frequency-based methods (e.g., Term Frequency-Inverse Document Frequency (TF-IDF) or Term Frequency (TF)) determine the importance of a word based on its frequency in a text. TF-IDF considers the frequency of a word in a text and its importance in the entire document set, whereas TF only considers the frequency of a word in a text.

[0168] Statistical methods in text analysis rely on statistical models to analyze the distribution and relationships of words in text. Common methods include Mutual Information, Pointwise Mutual Information, and Chi-square test. These methods usually require establishing a statistical model between words and text, and calculating the importance of words based on the model.

[0169] Text vectorization methods convert text into a vector representation and calculate the importance of words using a vector space model (VSM). Common methods include text vectorization methods such as Bag-of-words model (BoW), word embedding, and TF-IDF vectorization.

[0170] Machine learning methods use machine learning algorithms to train models that learn the importance of words from text. Common methods include text classification, text clustering, and keyword extraction models. These methods require model training using marked text data.

[0171] The above-described methods may be implemented by scripting in a programming language such as Python®.

[0172] The search module 332 includes a text search means 3111, a graphic search means 311, a conversion means 317, and a graphic matching means 318. The graphic search means 311 receives a graphic input by a user and searches the database module 600 to search for matching trademark cases. When the graphic search means 311 receives the input graphic, it first converts the graphic by the conversion means 317, adds Vienna classification tags, and converts the graphic into a vector. Then, the graphic matching means 318 performs a comparison based on the database module 600 to search for matching trademark cases. It also extracts past cases with a similarity score higher than a certain risk threshold.

[0173] The classification analysis module 333 may analyze the industry category codes of the input text by the classification module 304. The classification analysis module 333 is connected to the database module 600 to determine and create at least one set of trademark classification codes, which may also be trademark categories.

[0174] The text creation module 334 may include a risk management module 306 and a search document creation module 310, and may further include a text creation means 3091 and a diagram creation means 309. The search document creation module 310 may be part of the search module 332 described above.

[0175] The risk management module 306 receives the conceptual ideas and creates text and / or graphics based on the semantic analysis results combined with the content learning module 305 .

[0176] The keyword expansion module 335 receives and expands the keywords created by the keyword extraction module 331, which means searching for similar words or synonyms by comparing with a glossary to generate additional synonyms.

[0177] The document classification module 336 may classify all case files created by the user, for example, by applying specific classification labels, classifying the created cases into disclosure documents or search documents, patents or trademarks, etc.

[0178] The document classification module 336 may also classify similar figures retrieved by the figure retrieval means 311 in the database module 600. The similar figures are classified based on their similarity levels. For example, figures with a similarity score higher than the first risk threshold are classified into a high risk area, figures between the first and second risk thresholds are classified into a medium risk area, and figures below the second risk threshold are classified into a low risk area.

[0179] The auto-writing module 337 may be integrated with the risk management module 306 and the content learning module 305 to automatically write and create application documentation.

[0180] The search report generation module 338 may alternatively be replaced by the search document generation module 310, the disclosure document generation means 307, the claim content generation means 308 or the diagram generation means 309 described above.

[0181] 8 to 10 are flow charts of an example of a method executed by the system. The method is a method in which the electronic device 100 is operated by a user, the processor 101 of the electronic device 100 is connected to the server 200 via a network interface controller, and an application 201 for recommended categories and risks is executed, the method including at least:

[0182] (1) a user logs in through the electronic device 100, performs identity authentication, and confirms the user's ID and user data;

[0183] (2) a user creates a case order by means of the electronic device 100 and selects a first target country, and information in the case order is periodically updated in the temporary memory 400;

[0184] (3) The user inputs explanatory text or a figure using the electronic device 100, converts the explanatory text into a character string, and performs marking to create character string information. The input language of the character string information is recorded in the temporary memory 400;

[0185] (4) The semantic analysis module performs a semantic analysis on the character string information, and the classification module classifies the character string information into trademark categories.

[0186] In step (4), further

[0187] (411) A step of analyzing and dividing the character string information by a semantic analysis module to generate a semantic analysis result;

[0188] (412) generating at least one set of trademark classification codes based on the semantic analysis results by the classification module 304;

[0189] (413) The category recommendation module 700 includes a step of combining with the semantic analysis module 303 and the classification module 304 to analyze string information having ambiguous meaning (or inaccurate description) into trademark category recommendation information, and finally create a recommended trademark category.

[0190] In step (3), further

[0191] (31) a language determination module determines whether an input language of the character string information is the same as an official language of a first target country;

[0192] (32) If so, perform a semantic analysis;

[0193] (33) if not, first, translate the string information into the official language of the first target country by a translation module, and then perform semantic analysis. After step (413), translate the created recommended trademark category into the input language of the string information.

[0194] After step (413),

[0195] (421) The graphic search means and text search means 3111 receives the text and / or graphic input by the user, searches in the database module 600, compares and searches past cases of the trademark, creates a search document, and further filters the past cases with high similarity;

[0196] (422) A graphic retrieval means receives an input graphic, and converts the graphic into a vector by a conversion means, adding Vienna classification labels;

[0197] (423) Then, the input figure is compared with past cases in the database module by the figure matching means, and a figure having a similarity score exceeding a predetermined value is selected as a past case of the trademark, and the text search means 3111 performs text matching in a similar manner.

[0198] (5) The risk management module 306 creates text and / or graphics based on past cases matched by the graphic search means 311 and the text search means 3111 in the database module 600 based on the learning of the content learning module 305 .

[0199] Furthermore,

[0200] (51) A user inputs a trademark name or a conceptual idea of ​​a pattern through an input module 302, a pattern creation means 309 analyzes the conceptual idea, combines it with a semantic analysis module to translate it into a pattern creation language, creates a pattern code corresponding to the conceptual idea, and then creates a pattern corresponding to the conceptual idea through a compilation means, and a pattern search means 311 compares the created pattern with past cases searched in the past during the pattern creation process, and ensures that the similarity of the created pattern is less than a predetermined value;

[0201] (52) The text creation means 3091 includes a step of combining with a semantic analysis module to analyze conceptual ideas and create a text which is a trademark name, and further includes a step of using a text search means 3111 to compare the created text with past cases searched in the past, and ensure that the similarity of the created text is less than a predetermined value.

[0202] (6) The application document creation means extracts the user's login authentication identification data as application data, or permits the user to create an application document by simply inputting the application data into fields.

[0203] Once created, the case order may be completed and the patent may be submitted for filing.

[0204] Also, after step (6),

[0205] (7) the user selecting a second target country;

[0206] (71) a language determination module determining whether the language of the application is the same as an official language of a second target country;

[0207] (72) otherwise, first translating the application document into the official language of the second target country by a translation module, and then adjusting the format of the translated normative document and the application document by an application document preparation means;

[0208] (73) If so, adjust the format as is. After the format has been adjusted, the conversion of the international patent application is completed and submitted by the law firm to the official of the second target country.

[0209] Once the formatting adjustments are complete, the international patent application is converted and filed in the system, and then handed over to a law firm for formal filing with the authorities of the second target country.

[0210] Finally, the technical features and achievable technical effects of the present invention are summarized as follows.

[0211] First, the trademark risk management system and method of the present invention solves a common difficulty faced by people in the intellectual property field, especially those who have no concept of protecting their own technology with intellectual property rights, by providing intelligent recommendations.

[0212] Secondly, the trademark risk management system and method of the present invention can provide a quick recommendation, risk system and method for those who have difficulty in deciding what category to apply for even after determining the intellectual property type, thereby saving more time and cost.

[0213] Thirdly, the trademark risk management system and method of the present invention solves the pattern of small and medium-sized enterprises and venture enterprises without intellectual property departments, and reduces the effort and time required for brand propagation, as well as the propagation and understanding costs during that time, thereby avoiding unproposed scenarios.

[0214] Fourth, the trademark risk management system and method of the present invention provides a complete system and implementation method from the early stage of technology branding, recommends intellectual property types, and finally creates recommended categories and risks, which greatly reduces the professional hurdles for individuals to apply for intellectual property protection for their own professional technologies.

Claims

1. A trademark risk management method, comprising: an electronic device operated by a user; a processor of the electronic device connected to a server via a network interface controller; and executing an application program to perform category recommendation and risk management, the method comprising: (S100) A user inputs text content through a user interface of an electronic device, and a processor executes an input module to input the text content; (S200) A processor executes a semantic analysis module in an application program to analyze text content; (S300) the semantic analysis module is further connected to the classification module, the search module and the database module, for classifying the text content according to industry technology, performing a matching search in the database module, and sending the matching data to the intellectual property information disclosure module; (S400) the intellectual property information disclosure module analyzes and consolidates the data and presents the information to the user via a user interface of the electronic device; (S500) The analyzed and summarized data is also sent to a category recommendation module, which classifies and summarizes the intellectual property types in the data, and displays the recommended intellectual property types in a user interface in order of ranking and applies the recommended intellectual property types; (S600) a user selects a trademark type via a user interface; (S700) A user inputs a brand description through a user interface, and a processor executes an input module to input the brand description; (S800) the user completes the login process through the login module, the processor executes the semantic analysis module, the search module performs a matching search on the analysis result in the database module, and sends the search result to the recommendation module to generate a recommended application trademark category; (S900) The processor executes a search module to perform a second matching search in the database module based on the recommended application trademark category, and the second search result is sent to a search document creation module to create a risk assessment report; Trademark risk management methods, including:

2. After step (S900), (S901) The trademark risk management method of claim 1, further comprising a step of recreating text or graphics based on similar text or graphics in the risk assessment report and the user's brainstorming concept information received by the input module.

3. A trademark risk management system for receiving a client that receives a user by operating an electronic device, A processor of the electronic device is connected to a server via a network interface controller and executes an application program for category recommendation and risk management; The system comprises at least an input module for receiving text content input by a user, converting the text content into a character string, performing a marking process, transmitting character string information, and recording an input language of the character string information in a temporary memory; a semantic analysis module that receives the character string information, analyzes and segments the character string information using a natural language database, and creates and transmits a semantic analysis result; a classification module that analyzes the industry category classification code used in the semantic analysis result, and is connected to the database module to determine and generate at least one set of industry classification codes; a search module that performs a matching search in the database module based on the at least one set of industry classification codes and generates data of the matching search result; an intellectual property information disclosure module that receives the data and further statistically analyzes the data to generate basic intellectual property information; a recommendation module that further receives the data, classifies intellectual property types in the data, and creates the intellectual property types recommended for filing; a login module for authenticating the user's identity by operating the electronic device; Including, A trademark risk management system in which, when the user selects a trademark from the intellectual property type recommended for application, the input module receives the brand description re-input by the user and completes identity authentication through the login module; the semantic analysis module receives string information related to the brand description, analyzes and divides it to create and send the analysis result of the technical description; the search module performs a matching search on the analysis result in the database module and sends the search result to the recommendation module to create a trademark category recommended for application; and the search module performs a second matching search in the database module based on the recommended application trademark category and sends the second search result to the search document creation module to create a risk assessment report.

4. The trademark risk management system of claim 1, further comprising a risk management module that recreates text or graphics based on similar text or graphics in the risk assessment report and the user's brainstorming concept information received by the input module.

5. A trademark risk management system for receiving a client that receives a user by operating an electronic device, A processor of the electronic device is connected to a server via a network interface controller and executes an application program for category recommendation and risk management; The system comprises at least a login module for authenticating an ID of the user by operating the electronic device; an order processing module that creates a new case order after a first target country is selected by a user and periodically updates the information in the case order in a temporary memory of the system; an input module for receiving a text or a figure input by a user of the instruction manual, converting the text of the instruction manual into a character string for marking processing, transmitting the character string information, and recording the input language of the character string information in a temporary memory; a semantic analysis module for receiving the string information, analyzing and segmenting the string information using a natural language database, and generating and transmitting a semantic analysis result; A classification module that analyzes trademark category classification codes according to the semantic analysis results, and is connected to the database module, and is used to determine and generate at least one set of trademark classification codes; A category recommendation module is a calculation model of a natural language model, a trademark classification table and detailed training, which analyzes the string information with ambiguous meaning or inaccurate description into trademark category recommendation information, and finally combines with a semantic analysis module and a classification module to generate a recommended application trademark category; a text search means for receiving the input text, searching past cases in the database module, and further ranking the similarity to generate a risk assessment report; a graphic search means for receiving the input graphic, searching past cases in the database module, a conversion means, and a graphic comparison means; a content learning module, which is a large language model, further including a mode learning means for learning corresponding trademark content from said database module for different trademark categories; A risk management module further including a text creation means and a pattern creation means, which recreates text and / or graphics based on the learning of the content learning module and on past cases matched by the graphics search means and the text search means in the database module; Including, The pattern creation means, after the user inputs a brainstorming concept such as a trademark name or a figure via the input module, combines with the semantic analysis module to analyze the brainstorming concept and translate it into a pattern creation language, and creates a pattern code corresponding to the brainstorming concept using the pattern creation language. Thereafter, a compiler means creates a recreated figure corresponding to the brainstorming concept. The figure search means, during the recreation of the figure, compares the similarity between the recreated figure and the past case with a past case searched in the past so that the similarity between the recreated figure and the past case is less than a predetermined value. The text creation means is combined with the semantic analysis module to analyze the brainstorming concept after the user inputs the brainstorming concept of the trademark name or figure via the input module, and then combined with the content learning module to recreate the text; and at the same time, the text search means compares the similarity between the recreated text and the past case with the past case searched in the past so that the similarity between the recreated text and the past case falls within the range of the predetermined value. Trademark risk management system.

6. A trademark risk management system as described in claim 5, wherein when the graphic search means receives the input graphic, it first converts the graphic into a vector representation using a conversion means, and then performs a matching search in the database module using the graphic comparison means to find past cases.

7. The trademark risk management system according to claim 5 , further comprising a language determination module for determining whether an input language of the string information is the same as an official language of the first target country.

8. 8. The trademark risk management system according to claim 7, wherein if the language determination module determines that the input language of the character string information is different from the official language of the first target country, the language determination module translates the character string information using a translation module, and also translates the language of the trademark figure in the final search document into the input language of the character string information using the translation module.

9. The application processing module further includes an application creation means for extracting the identity information of the authenticated user through the login authentication and incorporating it into the application data, or for allowing the user to directly input the application data into the fields through the electronic device, receiving the application data through the input module, and creating the application data by applying a format template; 6. The trademark risk management system of claim 5, wherein after the case processing module creates the application, the user's case order is completed.

10. 6. A trademark risk management system according to claim 5, wherein said conversion means converts said graphic from pixels to vectors.

11. The trademark risk management system of claim 10, wherein the graphic comparison means calculates the similarity using edit distance, cosine similarity, or Jaccard similarity during the comparison process, and filters out graphics whose similarity exceeds a predetermined value.

12. The trademark risk management system of claim 5, further comprising a keyword extraction module that extracts keywords from the input content, creates a plurality of keywords, and uses the keywords to train the model with a machine learning algorithm.

13. 9. The trademark risk management system of claim 8, further comprising a multi-country conversion module, wherein after a second target country is selected by the user, the language determination module first determines whether the input language of the string information is the same as the official language of the second target country, and if not, translates the trademark name into the official language of the second target country before being formatted by the translation module.

14. (1) a user logs into an electronic device to authenticate the user's ID and identification information; (2) a user creates a case order by an electronic device and selects a first target country, whereupon information in the case order is periodically updated in a temporary storage; (3) a user inputs explanatory text or a graphic through an input module of the electronic device, converts the explanatory text into a character string, performs a marking process to create character string information, and records the input language of the character string information in a temporary memory; (4) a semantic analysis module performs a semantic analysis on the character string information, and a classification module performs a trademark classification on the character string information; Including, In step 4, further (411) A semantic analysis module analyzes and divides character string information to generate a semantic analysis result; (412) A classification module classifies the string information based on the semantic analysis result to generate at least one set of trademark classification codes; (413) A category recommendation module analyzes the string information into trademark category recommendation information, and combines it with a semantic analysis module and a classification module to finally create a recommended application trademark category; (421) The graphic search means and the text search means include a step of receiving the text and / or graphic input by the user, searching in the database module to compare and search past cases of the trademark, creating a search file, and further filtering past cases whose similarity is higher than the risk value; In step 421, further (5) The risk management module includes a step of recreating text and / or graphics based on past cases matched by the graphic search means and the text search means in the database module based on learning by the content learning module; How to manage trademark risk.

15. In step (3), further (31) a language determination module determines whether an input language of the character string information is the same as an official language of a first target country; (32) if so, performing semantic analysis directly; (33) if not, first translating the string information into the official language of the first target country by a translation module, and then performing semantic analysis; 15. The trademark risk management method according to claim 14, wherein in step (413), the created recommended application trademark category is translated again into an input language of the character string information.

16. After step (421), (422) The graphic retrieval means, after receiving the input graphic, converts the input graphic into a vector representation; (423) Then, the figure comparison means compares the figures in the database module, filters figures whose similarity exceeds a predetermined value, and the filtered figures are regarded as past examples of the trademark; 15. The trademark risk management method of claim 14, comprising:

17. In step (5), further (51) The graphic creation means, after a user inputs a brainstorming concept of a trademark name or a graphic through an input module, analyzes the brainstorming concept and combines with a semantic analysis module to translate the brainstorming concept into a graphic creation language, the graphic creation language creates a graphic code corresponding to the brainstorming concept, and then the graphic code is compiled by a compiler means to create a recreated graphic corresponding to the brainstorming concept; the graphic search means, in the process of recreating the graphic, compares the recreated graphic with a past case searched in the past to ensure that the recreated graphic and the past case have a similarity lower than a predetermined value; (52) The text creation means analyzes the concept of the brainstorm, and then combines with the semantic analysis module to recreate the text based on the content learning module, and the text creation means further compares the recreated text with past cases searched in the past to ensure that the recreated text and the past cases have a similarity lower than a predetermined value; 15. The trademark risk management method according to claim 14, comprising:

Images