Scientific and technological achievement transformation-oriented comprehensive consultation service and project management platform

Abstract

The invention discloses a comprehensive consultation service and project management platform oriented to scientific and technological achievement conversion, and relates to the technical field of scientific and technological achievement conversion consultation, and the platform comprises the steps: starting an achievement information registration and metadata extraction program for scientific and technological achievement information uploaded for the first time, and automatically generating a unique and non-repeatable achievement metadata tag; and storing the achievement information and the label in an achievement database in an associated manner to form an initial achievement entity model taking the label as a core. Uniform identification and centralized management of the achievements are achieved by distributing unique metadata labels for the achievements uploaded for the first time and establishing the main achievement entity model, and fragmentation of achievement information is avoided; meanwhile, by means of a natural language recognition and semantic aggregation algorithm, the platform can perform intelligent comparison and fine-grained combination on the subsequently uploaded result information, so that valuable supplementary information uploaded by each team is reserved, conflict fields generated due to inconsistent description are eliminated, and the integrity and consistency of the result content are ensured.

Description

Technical Field

[0001] The present invention relates to the technical field of transformation consulting of scientific and technological achievements, and particularly to an integrated consulting service and project management platform for scientific and technological achievement transformation. Background Art

[0002] An integrated consulting service and project management platform for scientific and technological achievement transformation is a digital system platform integrating scientific and technological achievement docking, transformation and implementation, project management and professional consulting services. By integrating the resources of universities, research institutions, enterprises and investment and financing institutions, the platform provides full-chain consulting services such as scientific and technological achievement information release, technology evaluation, market analysis, policy interpretation, intellectual property service, business model design, investment and financing docking, etc. At the same time, it also has the function of full life cycle management of projects, including project establishment, progress control, resource coordination, risk assessment and achievement evaluation, aiming to improve the efficiency of the transformation of scientific and technological achievements from the laboratory to industrialization and promote the deep integration of science and technology and the economy.

[0003] The existing technology has the following deficiencies:

[0004] Under the existing technology environment, there is still room for improvement in the management of scientific and technological achievements in universities. If the same scientific and technological achievement is uploaded to the platform by different team members or project team members respectively, thus forming multiple "similar but inconsistent in content" versions of achievement information, enterprises may misselect a version with a lower technology maturity or incomplete data for cooperation due to information confusion during docking. Such situations not only affect the efficient transformation of achievements, but also seriously weaken the credibility and professional image of universities and the platform.

[0005] The above information disclosed in the background art section is only used to enhance the understanding of the background of the present disclosure, and thus it may include information that does not constitute the prior art known to those of ordinary skill in the art. Summary of the Invention

[0006] The object of the present invention is to provide an integrated consulting service and project management platform for scientific and technological achievement transformation. By assigning a unique metadata tag to the first uploaded achievement and establishing a main achievement entity model, the unified identification and centralized management of achievements are realized, avoiding the fragmentation of achievement information; at the same time, with the help of natural language recognition and semantic aggregation algorithms, the platform can perform intelligent comparison and fine-grained merging of the subsequently uploaded achievement information, retaining the valuable supplementary information uploaded by each team and eliminating the conflicting fields caused by inconsistent descriptions, ensuring the integrity and consistency of the achievement content, so as to solve the problems in the above background art.

[0007] To achieve the above object, the present invention provides the following technical solutions: an integrated consulting service and project management platform for the transformation of scientific and technological achievements, including an achievement filing and identification module, a semantic parsing and feature extraction module, an achievement similarity determination and attribution judgment module, an achievement aggregation and version fusion module, and a version management and permission review module;

[0008] The achievement filing and identification module starts the registration and metadata extraction program of the scientific and technological achievement information for the first uploaded scientific and technological achievement information, automatically generates a unique and non-repeating achievement metadata label, and associates and stores the achievement information and the label in the achievement database to form an initial achievement entity model centered on the label;

[0009] The semantic parsing and feature extraction module performs text preprocessing on the subsequently uploaded scientific research achievement information, uses a pre-trained natural language recognition model to convert the newly uploaded achievement information into feature vectors, and compares and analyzes them one by one with the feature vectors of all scientific and technological achievements stored in the achievement database to calculate the semantic similarity value between the two;

[0010] The achievement similarity determination and attribution judgment module sets several similarity determination thresholds according to the obtained semantic similarity value, specifically including a high threshold range, a medium threshold range, and a low threshold range, to determine whether the subsequently uploaded scientific and technological achievement and the first uploaded scientific and technological achievement belong to the same scientific and technological achievement;

[0011] For the case where the achievements are identified as the same scientific and technological achievement, the achievement aggregation and version fusion module starts the automatic aggregation and version information fusion program, and based on the version information of the achievement first uploaded, uses the natural language recognition and semantic aggregation algorithm to perform a fine-grained comparison at the field level between the subsequently uploaded scientific and technological achievement and the first uploaded achievement; for the case where the achievement is identified as a new scientific and technological achievement, it is added, and a new metadata label is assigned;

[0012] The version management and permission review module generates a new version number for the scientific and technological achievement after the automatic information merger is completed, records the difference information and the corresponding modification records between all versions, and forms a complete version tracking and historical record system for the achievement entity model; at the same time, a unique achievement manager is designated for each achievement entity, and the permission to perform the final confirmation and review of the key fields is given to it.

[0013] Preferably, the achievement metadata label is one of the achievement unique identification code, the achievement digital object unique identifier, and the achievement semantic fingerprint code.

[0014] Preferably, the calculation steps of the semantic similarity value are as follows:

[0015] Call the pre-trained natural language recognition model to perform word segmentation, word vector conversion, and context semantic modeling on the newly uploaded text information, and output a feature vector u with a dimension of D, where D represents the vector dimension produced by the pre-trained language model;

[0016] For each piece of achievement information stored in the achievement database, call the same or compatible semantic model to obtain their respective feature vectors v i , where i represents the number of each record in the achievement database;

[0017] After generating the vectors, perform alignment processing on all vectors in terms of dimensions.

[0018] Preferably, after obtaining and aligning the feature vectors, use the following similarity formula to perform exponentially weighted calculation of the per-dimension differences between the newly uploaded achievement vector u=(u1,...,u D ) and the i-th achievement vector v i =(v i1 ,...,v iD ) in the database, and obtain a semantic similarity value:

[0019]

[0020] , where: S im (u, v i ) represents the semantic similarity value, α is the sensitivity adjustment parameter, α>0, used to control the influence weight of local dimension differences in the overall similarity; β is the smoothing factor, β>0; represents the weighted accumulation of the relative differences of each vector dimension to form the final distance metric and then map it to the similarity range of (0,1] through the exponential function. u k represents the k-th component of the semantic vector u corresponding to the newly uploaded scientific and technological achievement text, and v ik represents the k-th component of the feature vector v of the i-th existing achievement in the achievement database i , and k represents the dimension index of the semantic vector, that is, the current calculation is the k-th dimension.

[0021] Preferably, compare the obtained semantic similarity value with the high threshold range, medium threshold range, and low threshold range for comparison and analysis to determine whether the first uploaded scientific and technological achievement and the subsequently uploaded scientific and technological achievements belong to the same scientific and technological achievement. The specific determination steps are as follows:

[0022] If the semantic similarity value is in the high threshold range, it is determined that the subsequently uploaded scientific and technological achievement belongs to the same scientific and technological achievement as the first uploaded scientific and technological achievement;

[0023] If the semantic similarity value is in the medium value range, it is determined that the subsequent uploaded scientific and technological achievements are to be determined compared with the first uploaded scientific and technological achievements, and it is prompted that the administrator or the achievement uploader conducts manual review and final confirmation;

[0024] If the semantic similarity value is in the low threshold range, it is determined that the subsequent uploaded scientific and technological achievements are not the same scientific and technological achievements as the first uploaded scientific and technological achievements, and the subsequent uploaded scientific and technological achievements are determined as new scientific and technological achievements and are newly added.

[0025] Preferably, based on the version information of the first upload of the achievement, the subsequent uploaded scientific and technological achievements are compared with the first uploaded achievement at the field level in a fine-grained manner using natural language recognition and semantic aggregation algorithms. The specific steps are as follows:

[0026] Perform field-level data extraction and standardization operations on the information of the first uploaded achievement and the information of the subsequent uploaded achievement;

[0027] Use a pre-trained language model to perform embedding processing on the text content of each field and convert it into a multi-dimensional semantic vector representation.

[0028] Preferably, after completing the semantic vectorization, for each corresponding field, use an improved semantic distance function to calculate the semantic similarity between the first uploaded version and the subsequent uploaded version, and measure the degree of semantic consistency of each field;

[0029] For fields with a similarity greater than or equal to the set threshold, they are regarded as semantically repeated or highly consistent and do not need to be modified; for fields with a similarity less than the threshold, they are identified as information differences or potential conflicts, and are classified as "incremental information" or "conflict information" and marked in the difference analysis table.

[0030] Preferably, for the fields identified as "incremental information", they are directly appended to the content structure of the corresponding fields in the main achievement entity, and are fused as "additional information", and the source and version number are marked in the database;

[0031] For the field content identified as having conflicts, preprocess it according to the preset strategy, generate a recommended retention value for each conflict field, and retain all historical content as a version record.

[0032] Preferably, a new version is generated based on the original main achievement entity model, the incremental information and the processed field content are uniformly summarized into the new version data packet, and at the same time the additional metadata is bound under this version model; if it is judged that there are high risks or keyword field conflicts, the "audit transfer mechanism" is automatically triggered, and the integration result is submitted to the "unique manager" of this achievement for manual review and final confirmation.

[0033] In the above technical solution, the technical effects and advantages provided by the present invention are as follows:

[0034] By assigning a unique metadata tag to the results uploaded for the first time and establishing a main result entity model, the present invention realizes the unified identification and centralized management of results, avoiding the fragmentation of result information; at the same time, with the help of natural language recognition and semantic aggregation algorithms, the platform can perform intelligent comparison and fine-grained merging on the result information uploaded subsequently, retaining both the valuable supplementary information uploaded by each team and eliminating the conflicting fields caused by inconsistent descriptions, ensuring the integrity and consistency of the result content. In addition, the introduction of the version management mechanism and the result manager system enables universities to conduct traceable and controllable dynamic supervision on the evolution process of each result, improving the accuracy, credibility and transformation efficiency of result display, and truly realizing the transformation from "extensive uploading" to "intelligent aggregation and accurate release", building a set of intelligent, structured and reliable new models for the management of scientific and technological achievements for universities, and significantly enhancing the professional image and external cooperation attractiveness of universities in the process of result transformation. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required to be used in the embodiments. Obviously, the drawings described below are only some embodiments recorded in the present invention. For those of ordinary skill in the art, other drawings can also be obtained based on these drawings.

[0036] Figure 1 It is a schematic diagram of the modules of the comprehensive consulting service and project management platform for scientific and technological achievement transformation of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0037] Now, the exemplary embodiments will be described more fully with reference to the accompanying drawings. However, the exemplary embodiments can be implemented in various forms and should not be construed as limited to the examples set forth herein; rather, these exemplary embodiments are provided so that the present disclosure will be more complete and comprehensive, and will fully convey the concept of the exemplary embodiments to those skilled in the art.

[0038] The present invention provides a comprehensive consulting service and project management platform for scientific and technological achievement transformation as Figure 1 shown, including:

[0039] In a university or a scientific and technological achievement management platform, when team members or research group members first upload information on a brand-new scientific and technological achievement, the system first starts the registration and metadata extraction process for the achievement information, specifically including the parsing and structured storage of key information fields such as the achievement name, the affiliated discipline field, the abstract of the research content, the information of the research members and the team, the intellectual property status (such as the patent or software copyright number), and the technology maturity level, etc.; on this basis, the system automatically generates a unique and non-repeating achievement metadata tag (i.e., the main achievement identification code), and associates and stores the above-mentioned achievement information with the tag in the achievement database to form an initial achievement entity model centered around this tag.

[0040] Through this step, it can be ensured that each scientific and technological achievement has a unique identification and tracking identifier, laying the foundation for the unified and standardized management of achievement information, and effectively avoiding the problem of information redundancy caused by the subsequent multiple uploads of achievements.

[0041] In a scientific and technological achievement management platform, to ensure that each achievement has a unique identifier, the system needs to automatically generate an achievement metadata tag for each achievement uploaded for the first time. This tag is not only used for the unique identification of the achievement but also serves as the core index for subsequent version management, data tracking, and information aggregation. The following are three common forms of achievement metadata tags, which are elaborated in detail respectively:

[0042] 1. Achievement Unique Identifier Code (Achievement UID, Unique Identifier Code);

[0043] The achievement unique identifier code is a string number generated based on an automatic numbering rule, with uniqueness and non-repeatability. It usually adopts the form of a combination of letters and numbers, for example: "GZ2025-TF000123". The abbreviation of the university (such as GZ as the university code), the year (2025), the achievement type identifier (TF: scientific and technological achievement), and the serial number can be embedded in it. Such tags are convenient for machine reading and sorting, and are suitable for the unified archiving and rapid indexing of a large amount of achievement information. Through this achievement UID, the system can quickly locate the achievement entity model and achieve a unique mapping in the database. It is the most basic and commonly used type of achievement metadata tag.

[0044] 2. Digital Object Identifier for Achievements (Achievement DOI, Digital Object Identifier);

[0045] DOI is an international digital identifier system commonly used for academic papers and data resources, and can also be used for the annotation of scientific and technological achievements. In the scientific and technological achievement management system, the platform can generate exclusive DOI numbers for key achievements, such as: "10.12345 / univ.tech.2025.008", which are used for external citation, achievement release and permanent link. DOI has global uniqueness, traceability and persistent stability, and is especially suitable for high-value achievements with clear achievement forms (such as data packets, patent reports, transformation reports, etc.). Through DOI, achievements can be tracked and cited across platforms, enhancing the dissemination power and authority of scientific and technological achievements, and is especially suitable for the external display of achievement transformation, literature citation and intellectual property tracking.

[0046] 3. Semantic Fingerprint Code of Achievements;

[0047] The semantic fingerprint code of achievements is a hash identification code generated based on the extraction of natural language features of the achievement content. For example, using information such as the technical name, keywords, and research abstract of the achievement, a unique code is generated through semantic vectorization and hash compression, such as: "SFP-B3F7C9D812". Such tags have "content relevance" and can be used as the "semantic signature" of the achievement content to help the platform quickly perform semantic matching, similarity recognition and classification aggregation in subsequent uploads. This encoding not only realizes unique identification, but also has strong content recognition ability, and is suitable for building an intelligent achievement comparison and clustering system. It is a commonly used auxiliary identification method in achievement semantic recognition and machine learning algorithms.

[0048] Through the generation and application of these three types of achievement metadata tags, the scientific and technological achievement management system can realize the unique identification, version tracking, semantic analysis and cross-platform citation of achievement information, providing accurate, efficient and intelligent basic support for the transformation of scientific and technological achievements in universities.

[0049] When other teams or members of the research group upload new scientific and technological achievements subsequently, the system first preprocesses the uploaded scientific research achievement information (such as achievement name, technical abstract, patent data, technical maturity, etc.); subsequently, the system uses a pre-trained natural language recognition model (such as BERT, RoBERTa or other semantic models) to transform the newly uploaded achievement information into feature vectors, and compares and analyzes them one by one with the feature vectors of all scientific and technological achievements stored in the achievement database to calculate the semantic similarity value between the two;

[0050] Through the above steps, the platform can automatically and real-time determine the preliminary similarity between the subsequently uploaded achievement information and the existing achievements on the platform, providing objective and scientific data support for further judgment, ensuring the accurate matching of subsequent achievements with the original achievements, and reducing errors or omissions caused by manual comparison.

[0051] Text preprocessing specifically includes natural language word segmentation, stop word removal, named entity recognition, and feature word vector extraction.

[0052] The calculation steps of the semantic similarity value are as follows:

[0053] First, call a pre-trained natural language recognition model (such as BERT or RoBERTa) to perform word segmentation, word vector conversion, and context semantic modeling on the newly uploaded text information, and finally output a feature vector u with a dimension of D. Here, D represents the vector dimension produced by the pre-trained language model; immediately afterwards, for each piece of achievement information stored in the achievement database, the system also calls the same or compatible semantic model to obtain their respective feature vectors v i , where i represents the number of each record in the achievement database. After generating the vectors, the system performs alignment processing on all vectors in terms of dimensions to ensure that they belong to the same model and the same dimension, guaranteeing the comparability and consistency of subsequent similarity calculations;

[0054] Through this step, the system not only completes the high-dimensional semantic representation of the newly input text, but also lays a vector alignment foundation for subsequent similarity analysis, enabling the feature vectors to be accurately compared in the same embedding space.

[0055] After obtaining and aligning the feature vectors, the system uses the following similarity formula to perform an exponentially weighted calculation of the per-dimensional differences between the newly uploaded achievement vector u = (u1,..., u D ) and the i-th achievement vector v i = (v i1 ,..., v iD ) in the database, and obtains a semantic similarity value:

[0056]

[0057] , where: S im (u, v i ) represents the semantic similarity value, α is the sensitivity adjustment parameter, α > 0, which is used to control the influence weight of local dimensional differences in the overall similarity, and the larger the value, the more sensitive to local differences; β is the smoothing factor, β > 0, which avoids unstable results caused by the denominator being zero or too small; represents the weighted accumulation of the relative differences of each vector dimension, forms the final distance metric, and then maps it to the similarity range of (0, 1] through the exponential function. u k represents the k-th component of the semantic vector u corresponding to the newly uploaded scientific and technological achievement text, and v ik represents the feature vector (semantic feature vector) v of the i-th existing achievement in the achievement database iThe kth component of , where k represents the dimension index of the semantic vector, that is, the kth dimension is currently being calculated;

[0058] Through the exponential weighting form of this step, the system can delicately capture subtle dimensional differences while taking into account overall semantic differences, achieve accurate similarity assessment of text vectors, and provide quantitative criteria for subsequent confirmation of whether two results belong to the same technology or similar fields.

[0059] The system sets several similarity determination thresholds based on the semantic similarity values ​​obtained from the above comparison, including a high threshold interval, a medium threshold interval and a low threshold interval, to determine whether the subsequently uploaded scientific and technological achievements belong to the same scientific and technological achievements as the first uploaded scientific and technological achievements;

[0060] The semantic similarity values ​​obtained by comparison are compared and analyzed with the high threshold interval, the medium threshold interval and the low threshold interval to determine whether the scientific and technological achievements uploaded for the first time and the scientific and technological achievements uploaded subsequently belong to the same scientific and technological achievements. The specific determination steps are as follows:

[0061] If the semantic similarity value is within the high threshold range, it is determined that the subsequently uploaded scientific and technological achievements are the same as the first uploaded scientific and technological achievements;

[0062] If the semantic similarity value is in the medium range, it is determined that the subsequently uploaded scientific and technological achievements are pending confirmation with the first uploaded scientific and technological achievements, and the administrator or the person who uploaded the achievements is prompted to conduct manual review and final confirmation;

[0063] If the semantic similarity value is in the low threshold range, it is determined that the subsequently uploaded scientific and technological achievements are not the same as the first uploaded scientific and technological achievements, and the subsequently uploaded scientific and technological achievements are determined to be new scientific and technological achievements and added.

[0064] For example: if the similarity exceeds 80%, it is determined to belong to the same scientific and technological achievement and the merging procedure should be initiated; if the similarity is between 60% and 80%, the system will prompt the administrator or the person who uploaded the achievement to conduct manual review and final confirmation; if the similarity is lower than 60%, it will be automatically identified as a new scientific and technological achievement and assigned a new metadata tag.

[0065] Through clear and precise division of similarity intervals, this step can achieve rapid and accurate attribution determination of newly uploaded achievements, minimize subjective factors in achievement classification, and greatly improve the efficiency and accuracy of scientific and technological achievement classification.

[0066] If the system identifies the same scientific and technological achievements, the system will start the automatic aggregation and version information fusion program. Based on the version information of the first uploaded achievements, the system will use natural language recognition and semantic aggregation algorithms to perform a fine-grained field-level comparison between the subsequently uploaded scientific and technological achievements and the first uploaded achievements.

[0067] During this process, the incremental information contained in the newly uploaded achievements (such as new technical parameters, experimental data, or newly applied patent numbers) is automatically supplemented and integrated into the original achievement information. For information with obvious differences or conflicts (such as technology maturity assessment, patent status), the system determines the ultimately adopted information through a preset conflict resolution mechanism (such as field priority scoring based on parameters such as upload time, uploader weight, and information source reliability). If the system cannot make an automatic decision, the conflict fields are marked and manually reviewed and confirmed by the designated sole achievement manager;

[0068] Based on the version information of the achievement uploaded for the first time, the system uses natural language recognition and semantic aggregation algorithms to perform fine-grained field-level comparison between the subsequently uploaded scientific and technological achievements and the achievements uploaded for the first time. The specific steps are as follows:

[0069] The system first performs field-level data extraction and standardization operations on the achievement information uploaded for the first time and the subsequently uploaded achievement information. Each field (such as achievement title, technical abstract, application scenario, patent information, technology maturity, core parameters, etc.) is separately extracted and uniformly cleaned of natural language, including removing redundant formats, standardizing terms, eliminating synonyms ambiguity, etc. Subsequently, the system uses pre-trained language models (such as BERT, RoBERTa, etc.) to perform embedding processing on the text content of each field, converting it into a multi-dimensional semantic vector representation, laying a unified vector space foundation for subsequent semantic comparison. The role of this step is to ensure that the system has the ability to handle language heterogeneity and semantic ambiguity at the field level and achieve more accurate semantic comparison.

[0070] After completing the semantic vectorization, the system calculates the semantic similarity between the first uploaded version and the subsequently uploaded version for each corresponding field. Usually, improved semantic distance functions (such as weighted cosine distance, normalized exponential weighted difference, etc.) are used to accurately measure the degree of semantic consistency of each field. For fields with a similarity greater than or equal to the set threshold (such as more than 95%), the system considers them semantically repeated or highly consistent and does not need to be modified; for fields with a similarity less than the threshold, the system identifies them as information differences or potential conflicts and classifies them as "incremental information" or "conflict information" and marks them in the difference analysis table. The core role of this step is to accurately capture information differences by comparing each field to identify which information in the subsequent version needs to be supplemented, retained, or further confirmed.

[0071] For the fields identified as "incremental information" in the previous step, the system directly appends them to the content structure of the corresponding fields in the main achievement entity, integrates them as "additional information", and marks the source and version number in the system database. For the field content identified as having conflicts, the system preprocesses it according to preset policies, such as sorting field weights (based on factors such as the role of the uploader, information integrity, upload time, etc.), credibility assessment (such as whether it is accompanied by supporting evidence such as patent numbers, experimental reports, etc.), etc., to generate a recommended retention value for each conflicting field, while retaining all historical content as a version record. The core purpose of this step is to achieve the structured aggregation of achievement content on the premise of ensuring no information loss, and complete part of the automatic conflict decision-making through rule guidance.

[0072] When the information integration processing at the field level is completed, the system will generate a new version (such as v2.0) based on the original main achievement entity model, summarize the incremental information and the processed field content into the new version data packet, and at the same time bind additional metadata such as change logs, difference records, and merge sources under this version model. If the system determines that there are high risks or keyword field conflicts (such as major differences in technology maturity and core invention points), it will automatically trigger the "audit transfer mechanism" and submit the integration result to the "sole administrator" of this achievement for manual review and final confirmation. The key to this step is to ensure that the updated version after the system aggregates the achievement information has the support of an audit mechanism, prevent incorrect information from going online automatically, and ensure the authority, unity, and accuracy of the display of achievement information.

[0073] Through the above four steps, the system realizes the full-chain processing logic from "identifying the same achievement" to "intelligent aggregation and prudent integration", and constructs a scientific and technological achievement version management system that is both efficient and automatic and has credibility guarantee.

[0074] By using intelligent natural language recognition and semantic aggregation algorithms, the system recognizes that the newly uploaded results and the existing results belong to the same scientific and technological achievements, and then performs detailed, field-by-field intelligent comparison and aggregation and merging of the information of the two versions. Based on the information of the first uploaded results, the system accurately identifies and extracts the incremental content of key fields such as new or more complete technical indicators, experimental data, patent information and application scenarios in subsequent versions, and integrates these more complete and richer information into the existing main result entity model, thereby generating a new version of the results with more comprehensive, accurate and consistent information. This step can also automatically detect information conflicts or differences between the new and old versions, mark them out for subsequent manual confirmation and processing by management personnel, ensure the uniformity, authority and accuracy of scientific and technological achievement information on the platform, prevent enterprises from mistakenly selecting versions for cooperation due to inconsistent information, effectively enhance the credibility and professional image of university scientific and technological achievements in external display, and provide solid data support and guarantee for efficient achievement transformation and precise docking with enterprises.

[0075] Through the above steps, it can be ensured that all version information of the same achievement can be automatically and intelligently aggregated, redundant and conflicting information can be eliminated, and the authority, consistency and integrity of the version information displayed to the outside world can be ensured.

[0076] After each automatic merger of scientific and technological achievements, the system generates a new version number (such as V2.0, V3.0), and records the difference information and corresponding modification records between all versions, forming a complete version tracking and historical record system for the achievement entity model; at the same time, the platform designates a unique achievement manager for each achievement entity, granting him the authority to make final confirmation and review of key fields to ensure the accuracy, objectivity and standardization of the achievement information released to the outside world; when subsequent achievement versions conflict or require major modifications, the system will automatically trigger the review process, and the unique achievement manager will conduct manual review and confirmation before publishing, and record the review process; in addition, the system supports the public display of the latest approved achievement entity model version information, so that enterprises and partners can conduct efficient and barrier-free achievement precise docking;

[0077] Through the strict management and review system of this step, the authority, credibility, and real-time update of the results information can be effectively guaranteed, which will significantly improve the overall credibility and professional image of universities and platforms, and promote the efficient and orderly transformation of scientific and technological achievements.

[0078] Through the technical solution of the above comprehensive consulting service and project management platform for the transformation of scientific and technological achievements, universities have significantly improved the standardization, uniqueness, and evolvability controllability of information in scientific and technological achievement management, effectively solving the problems of information confusion and version conflicts caused by multiple team members uploading the same achievement repeatedly. The system realizes the unified identification and centralized management of achievements by assigning unique metadata tags to the achievements uploaded for the first time and establishing a main achievement entity model, avoiding the fragmentation of achievement information. At the same time, with the help of natural language recognition and semantic aggregation algorithms, the platform can perform intelligent comparison and fine-grained merging of the achievement information uploaded subsequently, retaining the valuable supplementary information uploaded by each team while eliminating the conflicting fields caused by inconsistent descriptions, ensuring the integrity and consistency of the achievement content. In addition, the introduction of the version management mechanism and the achievement manager system enables universities to conduct traceable and controllable dynamic supervision of the evolution process of each achievement, improving the accuracy, credibility, and transformation efficiency of achievement display, truly realizing the transformation from "extensive uploading" to "intelligent aggregation and precise release", building a set of intelligent, structured, and trustworthy new models for scientific and technological achievement management for universities, and significantly enhancing the professional image and external cooperation attractiveness of universities in the process of achievement transformation.

[0079] Only some exemplary embodiments of the present invention have been described above by way of illustration. Undoubtedly, for those of ordinary skill in the art, the described embodiments can be modified in various different ways without departing from the spirit and scope of the present invention. Therefore, the above drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

Claims

1. A comprehensive consulting service and project management platform for the transformation of scientific and technological achievements, characterized in that It includes an achievement filing and identification module, a semantic parsing and feature extraction module, an achievement similarity determination and attribution judgment module, an achievement aggregation and version fusion module, and a version management and permission review module; The achievement filing and identification module starts the registration and metadata extraction process of the achievement information for the first uploaded scientific and technological achievement information, automatically generates a unique and non-repeating achievement metadata label, and associates and stores the achievement information and the label in the achievement database to form an initial achievement entity model centered on the label; The semantic parsing and feature extraction module performs text preprocessing on the subsequently uploaded scientific research achievement information, uses a pre-trained natural language recognition model to convert the newly uploaded achievement information into a feature vector, and compares and analyzes it one by one with the feature vectors of all scientific and technological achievements stored in the achievement database to calculate the semantic similarity value between the two; The achievement similarity determination and attribution judgment module sets several similarity determination thresholds according to the obtained semantic similarity value, specifically including a high threshold range, a medium threshold range, and a low threshold range, to determine whether the subsequently uploaded scientific and technological achievement and the first uploaded scientific and technological achievement belong to the same scientific and technological achievement; For the case where it is identified as the same scientific and technological achievement, the achievement aggregation and version fusion module starts the automatic aggregation and version information fusion process, and based on the version information of the first uploaded achievement, uses the natural language recognition and semantic aggregation algorithm to perform a fine-grained comparison at the field level between the subsequently uploaded scientific and technological achievement and the first uploaded achievement; for the case where it is identified as a new scientific and technological achievement, it is added and a new metadata label is assigned; The version management and permission review module generates a new version number for the scientific and technological achievement after the automatic information merger is completed, records the difference information and the corresponding modification records between all versions, and forms a complete version tracking and historical record system for the achievement entity model; at the same time, it designates a unique achievement manager for each achievement entity and grants them the permission to perform the final confirmation and review of the key fields.

2. The comprehensive consulting service and project management platform for scientific and technological achievement transformation according to claim 1, wherein The achievement metadata label is one of the achievement unique identification code, the achievement digital object unique identifier, and the achievement semantic fingerprint code.

3. The comprehensive consulting service and project management platform for the transformation of scientific and technological achievements according to claim 1, characterized in that The calculation steps of the semantic similarity value are as follows: Call the pre-trained natural language recognition model to perform word segmentation, word vector conversion, and context semantic modeling on the newly uploaded text information, and output a feature vector u with a dimension of D, where D represents the vector dimension output by the pre-trained language model; For each piece of achievement information stored in the achievement database, call the same or compatible semantic model to obtain their respective feature vectors v i , where i represents the number of each record in the achievement database; After generating the vectors, perform alignment processing on all vectors in terms of dimensions.

4. The comprehensive consulting service and project management platform for the transformation of scientific and technological achievements according to claim 3, characterized in that, After obtaining and aligning the feature vectors, using the following similarity formula, for the newly uploaded result vector u = (u1,..., u D ), calculate the exponentially weighted differences dimension by dimension with the i-th result vector v i = (v i1 ,..., v iD ) in the database, and obtain a semantic similarity value: ， Where: S im (u, v i ) represents the semantic similarity value, α is the sensitivity adjustment parameter, α > 0, which is used to control the influence weight of the local dimension difference in the overall similarity; β is the smoothing factor, β > 0; represents weighted accumulation of the relative differences for each vector dimension to form the final distance metric, which is then mapped to the similarity range of (0, 1] through the exponential function. u k represents the k-th component of the semantic vector u corresponding to the newly uploaded scientific and technological achievement text, and v ik represents the k-th component of the feature vector v of the i-th existing achievement in the achievement database i , and k represents the dimension index of the semantic vector, that is, the current calculation is for the k-th dimension.

5. The comprehensive consulting service and project management platform for the transformation of scientific and technological achievements according to claim 4, characterized in that, Compare and analyze the obtained semantic similarity value with the high threshold range, the medium threshold range, and the low threshold range to determine whether the first uploaded scientific and technological achievement and the subsequently uploaded scientific and technological achievement belong to the same scientific and technological achievement. The specific determination steps are as follows: If the semantic similarity value is in the high threshold range, it is determined that the subsequently uploaded scientific and technological achievement and the first uploaded scientific and technological achievement belong to the same scientific and technological achievement; If the semantic similarity value is in the medium value range, it is determined that the subsequently uploaded scientific and technological achievement and the first uploaded scientific and technological achievement are to be determined, and it is prompted that the administrator or the achievement uploader perform manual review and final confirmation; If the semantic similarity value is in the low threshold range, it is determined that the subsequent uploaded scientific and technological achievements are not the same as the first uploaded scientific and technological achievements. The subsequent uploaded scientific and technological achievements are determined as new scientific and technological achievements and are newly added.

6. The comprehensive consulting service and project management platform for the transformation of scientific and technological achievements according to claim 1, characterized in that, Based on the version information of the first upload of the achievement, use natural language recognition and semantic aggregation algorithms to perform a fine-grained comparison of the fields between the subsequent uploaded scientific and technological achievements and the first uploaded achievements. The specific steps are as follows: Perform field-level data extraction and standardization operations on the achievement information of the first upload and the subsequent upload; Use a pre-trained language model to perform embedding processing on the text content of each field and convert it into a multi-dimensional semantic vector representation.

7. The comprehensive consulting service and project management platform for the transformation of scientific and technological achievements according to claim 6, characterized in that After completing the semantic vectorization, for each corresponding field, use an improved semantic distance function to calculate the semantic similarity between the first upload version and the subsequent upload version, and measure the degree of semantic consistency of each field; For fields with a similarity greater than or equal to the set threshold, they are regarded as semantically repeated or highly consistent and do not need to be modified; for fields with a similarity less than the threshold, they are identified as information differences or potential conflicts and are classified into "incremental information" or "conflict information" and marked in the difference analysis table.

8. The comprehensive consulting service and project management platform for the transformation of scientific and technological achievements according to claim 7, characterized in that, For the fields identified as "incremental information", directly append them to the content structure of the corresponding field in the main achievement entity, perform fusion processing as "additional information", and mark the source and version number in the database; For the field content identified as having conflicts, perform preprocessing according to the preset strategy, generate a recommended retention value for each conflict field, and retain all historical content as a version record.

9. The comprehensive consulting service and project management platform for the transformation of scientific and technological achievements according to claim 6, characterized in that Generate a new version based on the original main achievement entity model, summarize the incremental information and the processed field content into the new version data packet, and bind the additional metadata under this version model; if it is determined that there are high risks or keyword field conflicts, the "audit transfer mechanism" will be automatically triggered, and the integration result will be submitted to the "sole manager" of this achievement for manual review and final confirmation.

Images