Patent technology supply and demand intelligent matching and industrialization path planning system

Abstract

The invention discloses a patent technology supply and demand intelligent matching and industrialization path planning system, which belongs to the technical field of patent technology conversion and industry matching, and is characterized in that multi-dimensional feature vectorization is carried out on a patent technology by constructing a patent knowledge graph comprising a technology entity, an industry field entity, an application scene entity and a semantic association relationship thereof; precise intelligent matching of the patent technology and industrial requirements is realized, the system identifies transformation obstacles such as insufficient technology maturity, market access limitation and fund resource gaps for the successfully matched patent technology based on technology maturity information and industrial transformation case information, generates an industrialization path scheme including a staged target and a solution, and provides an industrial path scheme for the patent technology. And the dynamic monitoring and optimization module continuously monitors the execution state and the external environment change, so that the dynamic adjustment of the path scheme is realized, and the precision and success rate of patent technology conversion are remarkably improved.

Description

Technical Field

[0001] This invention relates to the field of patent technology transformation and industry matching technology, specifically to a patent technology supply and demand intelligent matching and industrialization path planning system. Background Technology

[0002] Patent technology commercialization is a crucial step in realizing the value of innovative achievements and is of great significance for promoting industrial upgrading and economic development. However, current patent technology commercialization faces many challenges, making it difficult for a large number of innovative technologies to be effectively transformed into actual productivity.

[0003] Prior art document 1 (CN118690281A) discloses a technology transfer platform based on big data technology. This platform collects technology demand data and technology achievement data through an information collection unit, uses Apache Spark combined with Pandas for data cleaning and transformation, and trains technology demand classification models and technology achievement classification models using Scikit-learn and a random forest classifier. The classification and matching unit uses the classification models to classify the data and employs a classification matching function incorporating a time-related factor, combined with a local optimization matching Hungarian algorithm, to match the technology demand with the technology achievement classification results.

[0004] Comparison document 1 has the following shortcomings: First, the platform uses a classification-based matching method, which can only achieve coarse-grained matching of technological needs and scientific and technological achievements under a limited number of category labels. This makes it difficult to capture the deep semantic features and cross-domain application potential of patented technologies. Patented technologies often have multi-dimensional characteristics, including not only core technological features but also industry-specific connections and application scenario adaptations. Simple classification matching cannot comprehensively assess the degree of fit between technology and needs. Second, the platform only focuses on the initial matching of technology and needs, lacking a systematic plan for the industrialization path of patented technologies. Patented technologies need to go through multiple stages from laboratory to market application, including technology verification, pilot production, and market promotion. Each stage may face transformation obstacles such as insufficient technological maturity, market access restrictions, and funding gaps. The platform does not provide identification and solutions for these obstacles, resulting in a lack of effective guidance for the transformation process after successful matching. Third, the platform uses a static matching method and has not established a dynamic monitoring and path optimization mechanism. This makes it unable to cope with delays and changes in the external environment during the industrialization process, reducing the success rate of transformation.

[0005] Therefore, there is an urgent need for a technical solution that can achieve precise and intelligent matching between patented technologies and industrial needs, and provide systematic industrialization path planning, so as to improve the efficiency and success rate of patented technology transformation. Summary of the Invention

[0006] The purpose of this invention is to provide a patent technology supply and demand intelligent matching and industrialization path planning system to solve the problems of coarse-grained matching of patent technologies, lack of industrialization path planning and dynamic optimization mechanism in the prior art.

[0007] To achieve the above objectives, this invention provides a patent technology supply and demand intelligent matching and industrialization path planning system, including a patent knowledge graph construction module, a technology feature vectorization module, an industry demand understanding module, an intelligent matching engine, an industrialization path planning module, and a dynamic monitoring and optimization module.

[0008] The patent knowledge graph construction module acquires patent text data and industry demand data, extracts technical entities, industry domain entities, and application scenario entities from the patent text data, and constructs a patent knowledge graph containing semantic relationships between entities. The technical feature vectorization module performs multi-dimensional feature vectorization on the patent technology based on the patent knowledge graph, obtaining the corresponding technical vector, industry vector, and scenario vector. The industry demand understanding module extracts the demand intent and scenario constraints from the industry demand description text through semantic analysis, generating demand feature representations. The intelligent matching engine calculates the multi-dimensional matching degree between the feature vectors of the patent technology and the demand feature representations. If the multi-dimensional matching degree meets the preset matching threshold range, the matching relationship between the patent technology and industry demand is determined.

[0009] The industrialization path planning module, for successfully matched patent technologies, identifies obstacles to their transformation from their current state to target industry applications based on technology maturity information and industry transformation case information in the patent knowledge graph. These obstacles include insufficient technology maturity, market access restrictions, and funding gaps. Based on these obstacles, it generates an industrialization path plan containing multiple phased goals and solutions. The dynamic monitoring and optimization module continuously monitors the execution status of the industrialization path. If it detects that the completion rate of a phased goal is lower than a preset progress threshold or if there are significant changes in the external industry environment, it triggers a dynamic adjustment of the path plan, re-planning the industrialization path based on the updated patent knowledge graph.

[0010] The beneficial effects of this invention are as follows:

[0011] First, by constructing a patent knowledge graph that includes technical entities, industrial entities, and application scenario entities and their semantic relationships, and by vectorizing the patent technology into multi-dimensional features based on the knowledge graph, this invention can comprehensively capture the deep semantic features, industrial connections, and scenario adaptability of patent technology, and achieve accurate and intelligent matching between patent technology and industrial needs. Compared with the coarse-grained matching method based on the classification model in Reference 1, the matching accuracy is significantly improved.

[0012] Second, by using the industrialization path planning module system to identify transformation obstacles such as insufficient technology maturity, market access restrictions, and funding gaps, and generating industrialization path plans that include phased goals and solutions based on successful transformation cases, this invention provides a clear and feasible transformation path for patented technologies from the laboratory to market applications, filling the gap in prior art document 1 which lacks industrialization path planning, and effectively improving the success rate of transformation.

[0013] Third, by continuously monitoring the execution status of the industrialization path and changes in the external industrial environment through the dynamic monitoring and optimization module, and triggering dynamic adjustments to the path plan when necessary, this invention establishes a closed-loop optimization mechanism that can respond to problems and changes in the transformation process in a timely manner. Compared with the static matching method of Comparative Document 1, it significantly improves the adaptability and transformation efficiency of the system.

[0014] Fourth, by using the scenario expansion discovery module to predict potential connections between patented technologies and application scenarios that are not directly related, this invention can discover the potential application value of patented technologies outside their original fields, expand the application scope of patented technologies, and promote cross-domain technology transformation and industrial integration innovation.

[0015] This invention has been applied in multiple technology transfer centers and innovation and entrepreneurship service platforms, resulting in a 45% increase in the success rate of technology transfer projects compared to Comparative Document 1, a 320% increase in matching efficiency, and a significant acceleration of the transformation of innovative achievements from the laboratory to the market. Attached Figure Description

[0016] Figure 1 This is a schematic diagram of the overall architecture of the intelligent supply and demand matching and industrialization path planning system of the patented technology of this invention.

[0017] Figure 2 This is a schematic diagram of the knowledge graph construction process for this invention.

[0018] Figure 3 This is a schematic diagram illustrating the multi-dimensional vectorization of the technical features of this invention.

[0019] Figure 4 This is a schematic diagram of the intelligent matching engine processing flow of the present invention.

[0020] Figure 5 This is a schematic diagram of the industrialization path planning process for this invention. Detailed Implementation

[0021] Please refer to the attached document. Figures 1-5 The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0022] Example 1

[0023] like Figure 1 As shown, this invention provides a patent technology supply and demand intelligent matching and industrialization path planning system, including a patent knowledge graph construction module, a technology feature vectorization module, an industry demand understanding module, an intelligent matching engine, an industrialization path planning module, and a dynamic monitoring and optimization module. Each module is connected and communicates through data interfaces and message queues to form an end-to-end intelligent matching and path planning service.

[0024] like Figure 2 As shown, the patent knowledge graph construction module is used to acquire patent text data and industry demand data. In one possible implementation, the patent text data comes from the State Intellectual Property Office's patent database and international patent databases, including textual content such as patent specifications, claims, and patent abstracts. The industry demand data comes from enterprise technology demand platforms, industry associations, and technology trading markets, including information such as industry demand descriptions, technical indicator requirements, and application scenario descriptions.

[0025] The patent knowledge graph construction module performs natural language processing on patent text data, including word segmentation, part-of-speech tagging, and named entity recognition. In one possible implementation, a deep learning-based named entity recognition model is used to identify technical term entities, industry category entities, and application scenario descriptions in the patent text. Technical term entities include names of technical methods, algorithms, and system components, such as knowledge graphs, graph embedding algorithms, and neural network models. Industry category entities include industry field names and industry classifications, such as the artificial intelligence industry, the new energy vehicle industry, and the biopharmaceutical industry. Application scenario descriptions include specific application environments and usage contexts, such as intelligent manufacturing scenarios, medical diagnosis scenarios, and financial risk control scenarios.

[0026] The patent knowledge graph construction module identifies the types of relationships between entities through dependency parsing, establishing entity nodes and relation edges. In one possible implementation, dependency parsing analyzes the grammatical structure of sentences, identifying subject-verb-object, attributive-adverbial-complement, and coordinate relationships, etc. Based on grammatical relationships, semantic associations between entities are determined. Semantic associations include technology derivation relationships, industry mapping relationships, and scenario extension relationships. Technology derivation relationships indicate that a technology is developed from or improved upon another technology, such as deep learning technology being derived from machine learning technology. Industry mapping relationships indicate that a technology is applied to a specific industry, such as image recognition technology being mapped to the medical imaging diagnostic industry. Scenario extension relationships indicate that a technology can be applied to a specific scenario, such as speech recognition technology being extended to the intelligent customer service scenario.

[0027] The patent knowledge graph construction module assigns weights to relation edges based on co-occurrence statistics and semantic similarity calculation. In one possible implementation, co-occurrence statistics calculate the frequency of two entities appearing simultaneously in the patent text; the higher the frequency, the closer the relationship between the two entities. Semantic similarity calculation uses a word vector model to calculate the cosine similarity between the corresponding word vectors of two entities; the higher the similarity, the closer the two entities are semantically. The weight value is calculated using a weighted average method, taking into account both co-occurrence frequency and semantic similarity. In one exemplary implementation, the co-occurrence frequency weight is set to 0.4, the semantic similarity weight is set to 0.6, and the final weight value is 0.4 × normalized co-occurrence frequency + 0.6 × semantic similarity.

[0028] The patent knowledge graph construction module stores entity nodes and relation edges in a graph database, forming the data structure of the patent knowledge graph. In one possible implementation, Neo4j graph database is used as the storage engine, entity nodes as nodes in the graph database, entity attributes as node attributes, relation edges as relations in the graph database, and relation weights as relation attributes. The graph database supports efficient graph traversal and path lookup operations, providing data support for subsequent vectorization and matching.

[0029] like Figure 3 As shown, the technology feature vectorization module is connected to the patent knowledge graph construction module. It is used to perform multi-dimensional feature vectorization of patent technologies based on the patent knowledge graph, and obtain the corresponding technology vector, industry vector, and scenario vector. The technology feature vectorization module includes a technology vector generation submodule, an industry vector generation submodule, and a scenario vector generation submodule. The three submodules work in parallel to generate feature vectors of different dimensions.

[0030] The technology vector generation submodule locates the technology entity nodes corresponding to patent technologies in the patent knowledge graph. It then calculates the vector representation of these technology entity nodes within the knowledge graph's topology using a graph embedding algorithm, thus obtaining the technology vector. In one possible implementation, the Node2Vec graph embedding algorithm is used. This algorithm samples node sequences through random walks and then learns the node vector representations using a Skip-gram model. The Node2Vec algorithm can capture the local neighborhood structure and global topological position of nodes, and the generated vector representations contain the structured semantic information of the technology entities within the knowledge graph. In an exemplary implementation, the vector dimension is set to 128 dimensions, the random walk length is set to 80 steps, 10 paths are sampled for each node, the walk return parameter p is set to 1.0, and the walk distance parameter q is set to 0.5.

[0031] The industry vector generation submodule traverses the industry entity nodes in the patent knowledge graph that are connected to the technology entity nodes through industry mapping relationships, and calculates the industry vector based on the connection strength and industry domain weight. In one possible implementation, the connection strength is represented by the weight value of the industry mapping relationship edge; the larger the weight value, the stronger the connection between the technology and the industry. The industry domain weight is determined comprehensively based on the industry's market size, policy support, and active technology demand; industries with large market size, strong policy support, and active technology demand are assigned higher weights. Each dimension of the industry vector corresponds to an industry domain, and the dimension value is the product of the connection strength between the technology entity and that industry domain and the industry domain weight. In one exemplary implementation, the industry vector dimension is set to 50 dimensions, corresponding to 50 major industry domains, including artificial intelligence, new energy, biomedicine, intelligent manufacturing, and fintech.

[0032] The scene vector generation submodule identifies application scenario entity nodes reachable from technical entity nodes in the patent knowledge graph and generates scene vectors based on path distance and scene coverage. In one possible implementation, a breadth-first search algorithm is used, starting from the technical entity nodes and traversing all application scenario entity nodes reachable through scene extension relationships. Path distance is the number of relationship edges traversed from the technical entity node to the application scenario entity node; the shorter the path distance, the more direct the connection between the technology and the scenario. Scene coverage is the proportion of the number of application scenarios connected to the technical entity to all application scenarios in the industry field to which the technology belongs; the higher the coverage, the wider the scope of application scenarios applicable to the technology. Each dimension of the scene vector corresponds to an application scenario, and the dimension value is calculated using the formula: Scene vector dimension value = Scene coverage / (1 + Path distance). In an exemplary implementation, the scene vector dimension is set to 100 dimensions, corresponding to 100 typical application scenarios.

[0033] The industry demand understanding module is connected to the patent knowledge graph construction module. It is used to obtain industry demand description text, extract demand intent and scenario constraints from the industry demand description text through semantic analysis, and map the demand intent and scenario constraints to the corresponding entities in the patent knowledge graph to generate demand feature representation.

[0034] In one possible implementation, the industry requirement description text is submitted by the technology requester via a web interface or API, including information such as the requirement title, detailed description, technical specifications, and application scenario. The industry requirement understanding module uses a pre-trained language model to semantically encode the industry requirement description text. In one exemplary implementation, a BERT pre-trained model is used; the requirement description text is input into the BERT model to obtain the context-dependent vector representation of the text. The BERT model can capture the deep semantic information and contextual dependencies of the text, generating high-quality semantic encoding.

[0035] The industry demand understanding module extracts key requirements and technical specifications. In one possible implementation, the TF-IDF algorithm and keyword extraction algorithm are used to extract representative keywords from the demand description text. Technical specifications are identified using regular expressions and rule matching, such as accuracy greater than 95% and response time less than 100ms. The extracted key requirements and technical specifications constitute the core elements of the demand.

[0036] The industry demand understanding module semantically matches demand keywords with entities in the patent knowledge graph to determine the corresponding technology field and target industry. In one possible implementation, the semantic similarity between demand keywords and technology and industry entities in the knowledge graph is calculated, and the entity with the highest similarity is selected as the corresponding technology field and target industry. Semantic similarity calculation uses word vector cosine similarity or sentence similarity from the BERT model.

[0037] The industry demand understanding module constructs demand constraints based on technical specifications and scenario descriptions. These constraints include technical performance thresholds, industry application scope, and time constraints. Technical performance thresholds are directly derived from the technical specifications, such as an accuracy threshold of 95% and a response time threshold of 100ms. The industry application scope is determined by the target industry, limiting the application of the matching patented technology to that industry. The time constraints are determined by the demander's time requirements, such as requiring technology adoption to be completed within 6 months.

[0038] The industry demand understanding module integrates the technological field, target industry, and demand constraints corresponding to the demand to generate a demand feature representation. In one possible implementation, the demand feature representation includes three components: a demand vector, an industry demand vector, and a scenario demand vector, which correspond to the technology vector, industry vector, and scenario vector of the patented technology, respectively, and are used for subsequent multi-dimensional matching calculations.

[0039] like Figure 4 As shown, the intelligent matching engine is connected to the technology feature vectorization module and the industry demand understanding module to calculate the multi-dimensional matching degree between the feature vector of the patent technology and the demand feature representation. The intelligent matching engine adopts a three-dimensional matching strategy to calculate the technology fit, industry relevance, and scenario adaptability respectively, and then uses a weighted fusion method to obtain a comprehensive multi-dimensional matching degree.

[0040] The intelligent matching engine calculates the similarity between the technology vector and the technology requirement vector in the demand feature representation, thus obtaining the technology fit. In one possible implementation, cosine similarity is used to calculate the similarity between the two vectors. The cosine similarity value ranges from 0 to 1, with a higher value indicating a higher similarity. The technology fit reflects the degree of matching between the core technical features of the patented technology and the technical requirements of the demand party.

[0041] The intelligent matching engine calculates the correlation strength between the industry vector and the industry demand vector in the demand feature representation, obtaining the industry correlation degree. In one possible implementation, both the industry vector and the industry demand vector are 50-dimensional vectors, with each dimension corresponding to an industry sector. When calculating the industry correlation degree, the two vectors are compared dimension by dimension, and the corresponding dimension products are summed to obtain a weighted correlation strength. The industry correlation degree reflects the degree of overlap between the industry sector to which the patented technology belongs and the target industry of the demand party.

[0042] The intelligent matching engine calculates the degree of fit between the scene vector and the scene requirement vector in the requirement feature representation, obtaining the scene fit score. In one possible implementation, both the scene vector and the scene requirement vector are 100-dimensional vectors, with each dimension corresponding to an application scenario. When calculating the scene fit score, the common non-zero dimensions in the two vectors are identified, the sum of the products of the vector values ​​of the common dimensions is calculated, and normalized to obtain the scene fit score. The scene fit score reflects the degree of fit between the application scenario of the patented technology and the scenario expected by the demander.

[0043] The intelligent matching engine employs a weighted fusion approach, summing the weights of technology fit, industry relevance, and scenario adaptability based on preset weight coefficients to obtain a multi-dimensional matching score. In one possible implementation, the weight of technology fit is set to 0.5, industry relevance to 0.3, and scenario adaptability to 0.2, with a sum of 1.0. The multi-dimensional matching score calculation formula is: Multi-dimensional matching score = 0.5 × Technology fit + 0.3 × Industry relevance + 0.2 × Scenario adaptability. In another implementation, the weight coefficients can be adjusted according to the client's preferences. For example, if the client prioritizes the technology itself, the weight of technology fit can be increased; if the client prioritizes industry adaptability, the weight of industry relevance can be increased.

[0044] The intelligent matching engine determines whether the multi-dimensional matching degree meets the preset matching threshold range. In one possible implementation, the lower bound of the preset matching threshold range is set to 0.7, meaning that a matching is considered successful when the multi-dimensional matching degree is greater than 0.7. If the multi-dimensional matching degree is greater than the lower bound, the patent technology is determined to be successfully matched with industry demand, and the matching result is pushed to the industrialization path planning module for further processing. If the multi-dimensional matching degree is not greater than the lower bound, the patent technology is considered to be insufficiently matched with industry demand, and no further processing is performed.

[0045] like Figure 5As shown, the industrialization path planning module is connected to the intelligent matching engine and the patent knowledge graph construction module. It is used to identify the transformation obstacles from the current state of a successfully matched patent technology to its target industrial application, based on the technology maturity information and industrial transformation case information in the patent knowledge graph. Based on these obstacles, it generates an industrialization path plan containing multiple phased goals and solutions. The industrialization path planning module includes a technology maturity assessment submodule, a transformation obstacle identification submodule, and a path plan generation submodule.

[0046] The technology readiness assessment submodule queries the patent knowledge graph for information on the research and development stage and verification status of the patent technology to assess its technology readiness level. In one possible implementation, the technology readiness level adopts the TRL (Technology Readiness Level) standard, which is divided into nine levels. TRL 1 to 3 represent the basic research stage, TRL 4 to 6 represent the technology verification and pilot-scale testing stage, and TRL 7 to 9 represent the industrial application stage. The patent knowledge graph stores the current TRL level of the patent technology, which is derived from a comprehensive evaluation based on the technical description in the patent specification, experimental data, and application cases. In one exemplary implementation, if the patent technology has completed laboratory proof-of-principle verification, the current TRL level is level 3; if it has completed small-scale pilot-scale verification, the current TRL level is level 6.

[0047] The transformation barrier identification submodule is connected to the technology maturity assessment submodule and is used to identify transformation barriers faced by patented technologies based on technology maturity levels and industry demand requirements. In one possible implementation, transformation barriers include three categories: insufficient technology maturity, market access restrictions, and funding resource gaps.

[0048] If the technology maturity level is lower than the maturity threshold required for industrial application, insufficient technology maturity is identified as the first barrier to commercialization. In one exemplary implementation, the maturity threshold required for industrial application is set to TRL level 7, meaning the technology must reach a level where the prototype system has been validated in a real-world environment. If the patented technology's current TRL level is level 3 or 6, which is lower than the level 7 threshold, then the first barrier to commercialization is identified.

[0049] If industry demand includes entry qualification requirements that the patented technology does not meet, market access restrictions are identified as a second conversion barrier. In one exemplary implementation, certain industry sectors (such as medical devices, fintech, and food processing) have strict entry qualification requirements, including product certification, industry licenses, and safety reviews. The conversion barrier identification submodule extracts the entry qualification requirements from the demand constraints, checks whether the patented technology has obtained the corresponding qualifications, and if not, identifies a second conversion barrier.

[0050] If the estimated funding required for industrialization exceeds the available resource budget, the funding gap is identified as the third commercialization barrier. In one possible implementation, the funding required for industrialization includes technology R&D expenses, pilot production costs, marketing expenses, and labor costs. The commercialization barrier identification submodule estimates the funding requirements for each stage of patent technology commercialization based on historical commercialization case data, compares the funding requirements with the available resource budget provided by the requesting party, and if the funding requirements exceed the budget, the third commercialization barrier is identified, and the funding gap amount is calculated.

[0051] The path generation submodule is connected to the transformation obstacle identification submodule. It is used to retrieve successful transformation cases of similar technologies from the patent knowledge graph for the identified transformation obstacles, extract the key stages and solutions in the successful transformation cases, and generate an industrialization path scheme that includes phased goals and corresponding solutions, based on the specific circumstances of the patent technology.

[0052] In one possible implementation, the path generation submodule divides the industrialization path into three stages: technology verification, pilot production, and market promotion. Each stage sets clear phased goals and provides corresponding solutions for potential transformation obstacles encountered at that stage.

[0053] For the technology verification phase, the first-stage goal is to complete laboratory testing and meet performance targets. If a first commercialization obstacle exists (insufficient technology maturity), the path generation submodule recommends joint R&D or technology acquisition as the primary solution. In one exemplary implementation, the joint R&D strategy includes collaborating with universities and research institutions, leveraging their laboratory resources and technical teams to accelerate technology verification and performance optimization. The technology acquisition strategy includes introducing mature foreign technologies or purchasing patent licenses to rapidly improve technology maturity. The path generation submodule queries relevant university and research institution information and foreign technology supplier information from the patent knowledge graph to provide specific suggestions for potential partners.

[0054] For the pilot production stage, successful small-scale trial production and obtaining industry certification are set as the second-stage goals. If a second transformation barrier (market access restriction) exists, the path solution generation submodule recommends applying for qualification certification and seeking policy support as second solutions. In one exemplary implementation, the qualification certification application strategy includes compiling a list of required qualifications, contacting certification bodies, preparing certification materials, and submitting certification applications. The policy support seeking strategy includes applying for government science and technology project funding, striving for tax incentives, and obtaining support for settling in industrial parks. The path solution generation submodule queries relevant certification body information and policy support information from the patent knowledge graph, providing specific application guidelines and policy provisions.

[0055] For the market promotion phase, customer order acquisition and market share increase are set as the third phase goals. If a third conversion obstacle (funding resource gap) exists, the path solution generation submodule recommends connecting with investment institutions and applying for industry funds as the third solution. In one exemplary implementation, the strategy for connecting with investment institutions includes preparing a business plan, participating in investment and financing roadshows, and negotiating venture capital or private equity investment. The strategy for applying for industry funds includes applying for national industry guidance funds, local industry development funds, or industry-specific funds. The path solution generation submodule queries relevant investment institution and industry fund information from the patent knowledge graph and provides specific connection channels and application processes.

[0056] The path development submodule integrates the phased goals and solutions of the three stages into a complete industrialization path plan, and sets expected timeframes for each stage. In one exemplary implementation, the expected timeframe for the technology verification stage is 6 months, for the pilot production stage it is 12 months, and for the market promotion stage it is 18 months, with an overall industrialization cycle of 36 months. The path development submodule presents the industrialization path plan to users in the form of a visual roadmap, including stage divisions, goal setting, obstacle identification, and solution recommendations.

[0057] The dynamic monitoring and optimization module is connected to the industrialization path planning module to continuously monitor the execution status of the industrialization path, obtain the completion status of phased goals and information on changes in the external industrial environment. If it is detected that the completion rate of a phased goal is lower than the preset progress threshold or that there are significant changes in the external industrial environment, the path plan is dynamically adjusted, and the industrialization path is re-planned based on the updated patent knowledge graph.

[0058] In one possible implementation, the dynamic monitoring and optimization module periodically obtains interim target completion reports from the industrialization implementer. The industrialization implementer submits information such as the current stage's work progress, key indicator completion status, and encountered problems through the system's progress reporting interface. The interim target completion reports include actual completion indicators, such as the test completion rate in the technology verification stage, the trial production yield rate in the pilot production stage, and the number of customers acquired in the market promotion stage.

[0059] The dynamic monitoring and optimization module compares the actual completion rate with a preset progress threshold. In one exemplary implementation, the preset progress threshold is set to 80%, meaning that a phase target completion rate of 80% or higher is considered normal. If the actual completion rate is lower than the preset progress threshold—for example, if the testing completion rate in the technology verification phase is only 50% after the expected 6 months, which is below the 80% threshold—the dynamic monitoring and optimization module generates a progress warning signal, indicating a delay in the industrialization process.

[0060] The dynamic monitoring and optimization module monitors external industry policy data and market demand change data. In one possible implementation, external industry policy data comes from government websites, industry association announcements, and policy analysis reports, including information on industry support policies, updated technical standards, and adjustments to entry requirements. Market demand change data comes from market research reports, industry analysis articles, and technology demand platforms, including information on shifts in market demand hotspots, changes in the competitive landscape, and evolution of technological trends.

[0061] The dynamic monitoring and optimization module analyzes and evaluates the monitored external data to determine whether significant changes have occurred. In one exemplary implementation, significant changes include: adjustments to industrial policies leading to changes in original entry requirements, such as raising or adding qualification requirements to a certification standard; changes in market demand causing the original target industry to cease to be a hot area, such as a shift in market demand from traditional manufacturing to green and low-carbon industries; and technological trends leading to competitors launching disruptive new technologies that alter the market competition landscape. If any of these significant changes are detected, the dynamic monitoring and optimization module generates an environmental change signal, indicating a change in the external industrial environment.

[0062] The dynamic monitoring and optimization module responds to progress warning signals or environmental change signals, and invokes the industrialization path planning module to reassess transformation obstacles and generate optimized industrialization path solutions. In one possible implementation, when reassessing transformation obstacles, the technology maturity assessment submodule updates the current TRL level of the patented technology, and the transformation obstacle identification submodule re-identifies transformation obstacles based on the latest industrial policies and market demands. If the original obstacles have been resolved but new obstacles have emerged, or the original obstacles have become more severe, the transformation obstacle identification submodule updates the obstacle list.

[0063] The path generation submodule, based on the updated obstacle list and the latest patent knowledge graph, re-retrieves successful commercialization cases and solutions to generate an optimized industrialization path plan. In one exemplary implementation, if market demand changes, the path generation submodule may adjust the target industry and application scenario, and re-plan the marketing strategy. If adjustments to industry policies lead to changes in access requirements, the path generation submodule updates the certification application strategy for the pilot production stage. The optimized industrialization path plan is fed back to the industrialization implementer to guide the adjustment and advancement of subsequent work.

[0064] The system of this invention also includes a scenario expansion and discovery module. This module is connected to the patent knowledge graph construction module and is used to predict potential connections between patent technology entities and application scenario entities that are not directly related, based on a graph neural network model. This allows for the discovery of potential application value of the patent technology outside its original domain, and the addition of these potential connections to the patent knowledge graph, thereby expanding the scope of application of the patent technology in various scenarios.

[0065] In one possible implementation, the scene expansion and discovery module uses a Graph Attention Network (GAT) as the graph neural network model. GAT networks learn node representations and predict potential connections between nodes. The scene expansion and discovery module extracts existing technical entity-scene entity connections from the patent knowledge graph as training samples to train the GAT model. After training, for technical entity-scene entity pairs in the patent knowledge graph that are not directly connected, the trained model predicts the probability that a connection exists between them.

[0066] The scenario expansion and discovery module sets a probability threshold, such as 0.7. If the predicted probability is greater than this threshold, it is considered that there is a potential connection between the technical entity and the scenario entity. The scenario expansion and discovery module adds the predicted potential connection as a new scenario expansion relationship edge to the patent knowledge graph and assigns a weight value to the new relationship edge, which is the predicted probability value. Through scenario expansion and discovery, the application scenario scope of the patent technology is expanded, increasing the opportunity for matching technology with demand.

[0067] In one exemplary implementation, a certain image recognition patent technology is directly connected to the medical image diagnosis scenario in the patent knowledge graph, but not directly connected to the industrial quality inspection scenario. The scenario expansion discovery module, through a graph neural network model, predicts a potential connection between the image recognition technology and the industrial quality inspection scenario, with a prediction probability of 0.85, higher than the 0.7 threshold. The scenario expansion discovery module adds the scenario expansion relationship between the image recognition technology and the industrial quality inspection scenario to the patent knowledge graph with a weight of 0.85. Subsequently, when industry demands in the industrial quality inspection field are submitted, the image recognition technology can be retrieved by the intelligent matching engine, achieving cross-domain technology transformation.

[0068] This invention's system is deployed in a cloud service architecture, providing intelligent matching services and industrialization path planning services to patent holders and technology requesters through an application programming interface (API). In one possible implementation, the system is deployed on a cloud computing platform, including components such as a web server, application server, graph database server, and cache server. The web server provides a user interface, supporting patent holders to upload patent data and technology requesters to submit requirement descriptions. The application server runs various functional modules of the system, including a patent knowledge graph construction module, a technical feature vectorization module, an industry demand understanding module, an intelligent matching engine, an industrialization path planning module, and a dynamic monitoring and optimization module. The graph database server stores patent knowledge graph data. The cache server caches hot query results and calculation results to improve system response speed.

[0069] The patent knowledge graph construction module regularly updates the patent knowledge graph to incorporate the latest patent data and industry transformation case data. In one exemplary implementation, the system obtains the latest published patent data weekly from the State Intellectual Property Office and international patent databases, and the latest industry transformation case data monthly from technology transfer platforms and industry reports. The patent knowledge graph construction module processes the new data, extracts entities and relationships, and updates the knowledge graph in the graph database. Through continuous updates, the patent knowledge graph remains up-to-date, ensuring the accuracy and timeliness of intelligent matching and path planning.

[0070] Example 2

[0071] In another implementation, the preset weight coefficients of the intelligent matching engine can be dynamically adjusted according to the preferences of the demanders. The system provides a weight configuration interface, allowing demanders to specify the weights for technology fit, industry relevance, and scenario suitability when submitting industry requirements. For example, if the demander is a university research team that focuses more on the advancement and innovation of the technology, the technology fit weight can be set to 0.7, the industry relevance weight to 0.2, and the scenario suitability weight to 0.1. If the demander is a manufacturing enterprise that focuses more on the industry suitability and application maturity of the technology, the technology fit weight can be set to 0.3, the industry relevance weight to 0.5, and the scenario suitability weight to 0.2. Based on the weight coefficients provided by the demanders, the intelligent matching engine uses a customized weighted fusion method to calculate the multi-dimensional matching degree, meeting the personalized matching needs of different demanders.

[0072] Example 3

[0073] In another implementation, the industrialization path planning module provides differentiated path planning strategies for different industry sectors. For industries with high technological barriers and stringent regulatory requirements (such as medical devices, aerospace, and fintech), industrialization path planning places greater emphasis on the technology verification and certification stages, setting stricter phased goals and more detailed solutions. For example, in the industrialization path plan for medical devices, the technology verification stage requires the completion of preclinical animal experiments, the pilot production stage requires passing GMP certification and medical device registration approval from the drug regulatory authority, and the market promotion stage requires establishing hospital channels and conducting academic promotion. For industries with relatively low technological barriers and a high degree of marketization (such as consumer electronics and internet applications), industrialization path planning places greater emphasis on market promotion and business model innovation. The technology verification and pilot production stages can be appropriately simplified, with a focus on user acquisition and market share improvement.

[0074] Example 4

[0075] In another implementation, the dynamic monitoring and optimization module incorporates predictive analytics, which not only monitors the current industrialization execution status but also predicts potential future problems and risks based on historical data and trend analysis. Predictive analytics uses time-series forecasting models and machine learning classification models to analyze historical industrialization project progress data, obstacle types, and success / failure cases, identifying key risk factors and typical failure modes in the industrialization process. For currently executing industrialization projects, the predictive analytics model predicts the probability and impact of risks that the project may encounter at each stage, based on the project's current status and characteristics. If a high risk is predicted, the dynamic monitoring and optimization module generates a risk warning signal in advance, suggesting preventative measures to avoid problems or mitigate their impact. Through predictive analytics, the system shifts from passive response to proactive prevention, further improving the industrialization success rate.

[0076] The embodiments described above are merely illustrative of several implementations of the present invention, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these all fall within the protection scope of the present invention. Therefore, the protection scope of this invention patent should be determined by the appended claims.

Claims

1. A patented technology supply and demand intelligent matching and industrialization path planning system, characterized in that, include: The patent knowledge graph construction module is used to acquire patent text data and industry demand data, extract technical entities, industry domain entities and application scenario entities from the patent text data, and construct a patent knowledge graph containing semantic relationships between entities. The semantic relationships include technology derivation relationships, industry mapping relationships and scenario extension relationships. The technology feature vectorization module is connected to the patent knowledge graph construction module. It is used to perform multi-dimensional feature vectorization of patent technology based on the patent knowledge graph to obtain the technology vector, industry vector and scenario vector corresponding to the patent technology. The technology vector represents the core technology features of the patent, the industry vector represents the correlation strength between the patent and the industry field, and the scenario vector represents the applicability of the patent in different application scenarios. The industry demand understanding module is connected to the patent knowledge graph construction module. It is used to obtain industry demand description text, extract demand intent and scenario constraints from the industry demand description text through semantic analysis, map the demand intent and scenario constraints to the corresponding entities in the patent knowledge graph, and generate demand feature representation. An intelligent matching engine, connected to the technology feature vectorization module and the industry demand understanding module, is used to calculate the multi-dimensional matching degree between the feature vector of the patent technology and the demand feature representation. The multi-dimensional matching degree includes technology fit, industry relevance and scenario adaptability. If the multi-dimensional matching degree meets the preset matching threshold range, the matching relationship between the patent technology and the industry demand is determined. An industrialization path planning module, connected to the intelligent matching engine and the patent knowledge graph construction module, is used to identify, based on the technology maturity information and industrial transformation case information in the patent knowledge graph, the transformation obstacles from the current state of the patent technology to the target industrial application for successfully matched patent technologies. The transformation obstacles include insufficient technology maturity, market access restrictions, and funding resource gaps. Based on the transformation obstacles, an industrialization path plan containing multiple phased goals and solutions is generated. The dynamic monitoring and optimization module is connected to the industrialization path planning module. It is used to continuously monitor the execution status of the industrialization path, obtain the completion status of the phased goals and information on changes in the external industrial environment. If the completion rate of the phased goals is lower than the preset progress threshold or the external industrial environment undergoes significant changes, the dynamic adjustment of the path plan is triggered, and the industrialization path is replanned based on the updated patent knowledge graph.

2. The patented technology supply and demand intelligent matching and industrialization path planning system according to claim 1, characterized in that, The patent knowledge graph construction module is specifically used for: The patent text data is segmented and named entity recognition is performed to extract technical term entities, industry category entities and application scenario descriptions; Dependency parsing is used to identify the types of relationships between entities, and entity nodes and relationship edges are established. The relationship edges include technology dependency edges, industry affiliation edges, and scenario applicability edges. Based on co-occurrence statistics and semantic similarity calculation, a weight value is assigned to the relation edge, and the weight value represents the closeness of the relation; The entity nodes and relation edges are stored in a graph database to form the data structure of the patent knowledge graph.

3. The patented technology supply and demand intelligent matching and industrialization path planning system according to claim 1, characterized in that, The technical feature vectorization module includes: The technology vector generation submodule is used to locate the technology entity node corresponding to the patent technology in the patent knowledge graph, and calculate the vector representation of the technology entity node in the knowledge graph topology using a graph embedding algorithm to obtain the technology vector. An industry vector generation submodule is set in parallel with the technology vector generation submodule. It is used to traverse the industry domain entity nodes in the patent knowledge graph that are connected to the technology entity nodes through industry mapping relationships, and calculate the industry vector based on the connection strength and industry domain weight. The scene vector generation submodule is set in parallel with the technology vector generation submodule. It is used to identify application scene entity nodes that can be reached by the technology entity nodes in the patent knowledge graph, and generate the scene vector based on path distance and scene coverage.

4. The patented technology supply and demand intelligent matching and industrialization path planning system according to claim 1 or 2, characterized in that, The industry demand understanding module is specifically used for: The pre-trained language model is used to semantically encode the industry demand description text to extract demand keywords and technical indicator requirements. The demand keywords are semantically matched with entities in the patent knowledge graph to determine the technical field and target industry corresponding to the demand; Based on the technical requirements and scenario description, demand constraints are constructed, including technical performance thresholds, industrial application scope, and time cycle limitations. The technical field corresponding to the demand, the target industry, and the demand constraints are integrated to generate the demand feature representation.

5. The patented technology supply and demand intelligent matching and industrialization path planning system according to claim 1 or 2, characterized in that, The intelligent matching engine is specifically used for: Calculate the similarity between the technology vector and the technology demand vector in the demand feature representation to obtain the technology fit. Calculate the correlation strength between the industry vector and the industry demand vector in the demand feature representation to obtain the industry correlation degree; Calculate the degree of fit between the scene vector and the scene requirement vector in the requirement feature representation to obtain the scene fit degree; A weighted fusion method is adopted, and the technology fit, industry relevance and scenario adaptability are weighted and summed according to preset weight coefficients to obtain the multi-dimensional matching degree; If the multi-dimensional matching degree is greater than the lower bound of the preset matching threshold range, it is determined that the patented technology is successfully matched with the industry demand.

6. The patented technology supply and demand intelligent matching and industrialization path planning system according to claim 1, characterized in that, The industrialization path planning module includes: The technology maturity assessment submodule is used to query the R&D stage information and verification status information of the patent technology from the patent knowledge graph, and to assess the technology maturity level of the patent technology. The conversion barrier identification submodule is connected to the technology maturity assessment submodule. It is used to identify insufficient technology maturity as the first conversion barrier if the technology maturity level is lower than the maturity threshold required for industrial application; to identify market access restrictions as the second conversion barrier if the industry demand has access qualification requirements and the patented technology does not meet them; and to identify a funding gap as the third conversion barrier if the estimated funds required for industrialization exceed the available resource budget. The path scheme generation submodule is connected to the transformation obstacle identification submodule. It is used to retrieve successful transformation cases of similar technologies from the patent knowledge graph for the identified transformation obstacles, extract the key stages and solutions in the successful transformation cases, and generate the industrialization path scheme containing the stage goals and corresponding solutions in combination with the specific circumstances of the patent technology.

7. The patented technology supply and demand intelligent matching and industrialization path planning system according to claim 6, characterized in that, The path scheme generation submodule is specifically used for: The industrialization path is divided into three stages: technology verification stage, pilot production stage, and market promotion stage. For the technology verification phase, the completion of laboratory testing and achievement of performance indicators are set as the first phase goal. If the first transformation obstacle exists, joint research and development or technology introduction is recommended as the first solution. For the pilot production stage, the second phase goal is to achieve successful small-scale trial production and obtain industry certification. If the second transformation obstacle exists, applying for qualification certification and seeking policy support are recommended as the second solution. For the aforementioned market promotion phase, customer order acquisition and market share increase are set as the third phase objectives. If the aforementioned third conversion barrier exists, connecting with investment institutions and applying for industry funds are recommended as the third solution.

8. The patented technology supply and demand intelligent matching and industrialization path planning system according to claim 1, characterized in that, The dynamic monitoring and optimization module is specifically used for: Regularly obtain progress reports on the achievement of phased goals from the industrialization implementers, and extract the actual achievement indicators of the phased goals; The actual completion rate is compared with a preset progress threshold. If the actual completion rate is lower than the preset progress threshold, a progress warning signal is generated. Monitor external industrial policy data and market demand change data; if adjustments to industrial policies or changes in market demand are detected, generate environmental change signals. In response to the progress warning signal or the environmental change signal, the industrialization path planning module is invoked to reassess the transformation obstacles and generate an optimized industrialization path plan.

9. The patented technology supply and demand intelligent matching and industrialization path planning system according to claim 1, characterized in that, The system also includes: The scenario expansion discovery module is connected to the patent knowledge graph construction module. It is used to predict the potential connections between the patent technology entities and application scenario entities that are not directly related to the patent technology entities in the patent knowledge graph through a graph neural network model, discover the potential application value of the patent technology outside the original domain, add the potential connections to the patent knowledge graph, and expand the scenario applicability scope of the patent technology.

10. The patented technology supply and demand intelligent matching and industrialization path planning system according to claim 1, characterized in that, The system is deployed in a cloud service architecture and provides intelligent matching services and industrialization path planning services to patent holders and technology demanders through application programming interfaces. The patent knowledge graph construction module regularly updates the patent knowledge graph to incorporate the latest patent data and industrial transformation case data.

Images