技术领域Technical Field
本发明涉及科技成果推荐技术领域，尤其涉及一种基于人工智能交互的科技成果推荐系统。The present invention relates to the technical field of scientific and technological achievement recommendation, and in particular to a scientific and technological achievement recommendation system based on artificial intelligence interaction.
背景技术Background Art
目前，科技活动的复杂性和不确定性逐渐加大，对科技成果推荐的准确性和实效性提出更高的要求。At present, the complexity and uncertainty of scientific and technological activities are gradually increasing, which puts higher requirements on the accuracy and effectiveness of the recommendation of scientific and technological achievements.
过于精准的科技成果推荐可能会导致信息茧房效应，严重限制了用户视野的拓展。因此，科技成果推荐系统如何利用在已知用户偏好进行精准推荐和探索用户未知兴趣之间找到平衡点，是目前科技成果推荐的难点所在。Too precise recommendations of scientific and technological achievements may lead to the information cocoon effect, which seriously limits the expansion of users' horizons. Therefore, how scientific and technological achievement recommendation systems can find a balance between making precise recommendations based on known user preferences and exploring users' unknown interests is the difficulty of scientific and technological achievement recommendation.
因此，如何结合具体的科技成果特性和用户需求设计科技成果推荐方式，是目前亟需解决的技术问题。Therefore, how to design a method to recommend scientific and technological achievements based on the specific characteristics of scientific and technological achievements and user needs is a technical problem that needs to be solved urgently.
发明内容Summary of the invention
鉴于上述的分析，本发明实施例旨在提供一种基于人工智能交互的科技成果推荐系统，用以解决现有技术中存在的过于精准的科技成果推荐可能会导致信息茧房效应的问题。In view of the above analysis, an embodiment of the present invention aims to provide a scientific and technological achievement recommendation system based on artificial intelligence interaction, so as to solve the problem in the prior art that overly precise scientific and technological achievement recommendations may lead to the information cocoon effect.
本发明为实现上述发明目的采用如下技术方案：The present invention adopts the following technical solutions to achieve the above-mentioned invention object:
一种基于人工智能交互的科技成果推荐系统，所述科技成果推荐系统包括用户输入模块、人工智能交互模块、关联分析模块、向量构建模块、科技成果预测模块及输出控制模块；其中，A scientific and technological achievement recommendation system based on artificial intelligence interaction, the scientific and technological achievement recommendation system comprises a user input module, an artificial intelligence interaction module, a correlation analysis module, a vector construction module, a scientific and technological achievement prediction module and an output control module; wherein,
用户输入模块，接收用户输入的科技成果需求信息；A user input module receives the scientific and technological achievement demand information input by the user;
人工智能交互模块，利用人工智能交互系统对科技成果需求信息进行智能检索，获得科技成果联合需求信息；Artificial intelligence interaction module, which uses the artificial intelligence interaction system to intelligently search for scientific and technological achievement demand information and obtain joint demand information for scientific and technological achievements;
向量构建模块，基于科技成果需求信息及科技成果联合需求信息，构建科技成果需求表征向量；Vector construction module, which constructs the scientific and technological achievement demand representation vector based on the scientific and technological achievement demand information and the scientific and technological achievement joint demand information;
科技成果预测模块，将科技成果需求表征向量输入科技成果预测模型，由科技成果预测模型预测输出参考科技成果；The scientific and technological achievement prediction module inputs the scientific and technological achievement demand representation vector into the scientific and technological achievement prediction model, and the scientific and technological achievement prediction model predicts and outputs reference scientific and technological achievements;
输出控制模块，将科技成果需求信息的关键特征矩阵与参考科技成果的关键特征矩阵进行匹配判断，从参考科技成果中选出匹配判断通过的科技成果推荐给用户。The output control module matches and judges the key feature matrix of the scientific and technological achievement demand information with the key feature matrix of the reference scientific and technological achievements, and selects the scientific and technological achievements that pass the matching judgment from the reference scientific and technological achievements and recommends them to the users.
所述科技成果推荐系统还包括可视化展示模块，用于将匹配判断通过的科技成果按照匹配度由高到低的顺序排序，进行可视化展示。The scientific and technological achievements recommendation system also includes a visual display module, which is used to sort the scientific and technological achievements that have passed the matching judgment in order from high to low matching degree for visual display.
所述用户输入模块内置用户输入模板，用户依据用户输入模板输入科技成果需求信息；其中，The user input module has a built-in user input template, and the user inputs the scientific and technological achievement demand information according to the user input template; wherein,
所述用户输入模板包括：科技成果需求的技术领域、技术要点及技术关键词。The user input template includes: technical fields, technical points and technical keywords required for scientific and technological achievements.
所述人工智能交互模块内置人工智能交互模板，用于将用户输入的科技成果需求信息置入人工智能交互模板，形成人工智能交互语句；由人工智能交互系统以人工智能交互语句为输入，对科技成果需求信息进行智能检索，获得科技成果联合需求信息。The artificial intelligence interaction module has a built-in artificial intelligence interaction template, which is used to place the scientific and technological achievement demand information input by the user into the artificial intelligence interaction template to form an artificial intelligence interaction statement; the artificial intelligence interaction system uses the artificial intelligence interaction statement as input to intelligently search the scientific and technological achievement demand information to obtain the joint demand information of scientific and technological achievements.
所述向量构建模块包括知识图谱构建单元和图谱表征单元；其中，The vector construction module includes a knowledge graph construction unit and a graph representation unit; wherein,
知识图谱构建单元，用于从科技成果需求信息及科技成果联合需求信息中提取出实体、实体属性及实体间的关系，并基于所述实体、实体属性及实体间的关系，构建知识图谱；A knowledge graph construction unit, used to extract entities, entity attributes and relationships between entities from the scientific and technological achievement demand information and the scientific and technological achievement joint demand information, and to construct a knowledge graph based on the entities, entity attributes and relationships between entities;
图谱表征单元，使用图嵌入算法对知识图谱进行图谱表征，构建科技成果需求表征向量。The graph representation unit uses a graph embedding algorithm to represent the knowledge graph and construct a representation vector for scientific and technological achievement requirements.
科技成果预测模块中的科技成果预测模型通过以下方式训练得到：The scientific and technological achievement prediction model in the scientific and technological achievement prediction module is trained in the following way:
获取多组训练样本数据，每组样本数据包括用户的科技成果需求信息样本、科技成果样本、以及各科技成果样本与相应用户的科技成果需求信息样本的匹配度得分；Acquire multiple sets of training sample data, each set of sample data includes a user's scientific and technological achievement demand information sample, a scientific and technological achievement sample, and a matching score between each scientific and technological achievement sample and the corresponding user's scientific and technological achievement demand information sample;
获取样本数据中用户的科技成果需求信息样本、科技成果样本的表征向量；将用户科技成果需求信息样本的表征向量输入科技成果预测模型，并用以训练拟合相应科技成果样本的表征向量的匹配度得分；Obtaining the user's scientific and technological achievement demand information sample and the representation vector of the scientific and technological achievement sample in the sample data; inputting the representation vector of the user's scientific and technological achievement demand information sample into the scientific and technological achievement prediction model, and using it to train and fit the matching score of the representation vector of the corresponding scientific and technological achievement sample;
直至训练完成，得到训练好的科技成果预测模型。Once the training is completed, a well-trained scientific and technological achievement prediction model is obtained.
科技成果预测模块，将科技成果需求表征向量输入科技成果预测模型后，由科技成果预测模型根据科技成果库中各科技成果表征向量与科技成果需求表征向量的匹配程度，预测输出参考科技成果。The scientific and technological achievement prediction module inputs the scientific and technological achievement demand representation vector into the scientific and technological achievement prediction model. The scientific and technological achievement prediction model then predicts and outputs reference scientific and technological achievements based on the degree of matching between the scientific and technological achievement representation vectors in the scientific and technological achievement database and the scientific and technological achievement demand representation vectors.
所述输出控制模块包括用户关键特征矩阵提取单元、匹配判断单元和筛选单元；其中，The output control module includes a user key feature matrix extraction unit, a matching judgment unit and a screening unit; wherein,
用户关键特征矩阵提取单元，用于从用户输入的科技成果需求信息中提取包含技术关键词的语句作为关键技术语句；并从关键技术语句中提取关键特征矩阵；A user key feature matrix extraction unit is used to extract sentences containing technical keywords as key technical sentences from the scientific and technological achievement demand information input by the user; and to extract a key feature matrix from the key technical sentences;
匹配判断单元，将科技成果需求信息的关键特征矩阵与参考科技成果的关键特征矩阵利用相似度原则进行匹配判断，得到各参考科技成果对应的匹配判断结果；The matching judgment unit matches and judges the key feature matrix of the scientific and technological achievement demand information with the key feature matrix of the reference scientific and technological achievements using the similarity principle to obtain the matching judgment results corresponding to each reference scientific and technological achievement;
筛选单元，根据预设匹配阈值，从参考科技成果中选出匹配判断结果超过匹配阈值的科技成果，并将其作为匹配判断通过的科技成果推荐给用户。The screening unit selects scientific and technological achievements whose matching judgment results exceed the matching threshold from the reference scientific and technological achievements according to the preset matching threshold, and recommends them to the user as scientific and technological achievements that pass the matching judgment.
从关键技术语句中提取关键特征矩阵，执行：Extract the key feature matrix from the key technical statements and execute:
对用户的科技成果需求信息对应的关键技术语句按照预设短语规则进行重构，得到相应关键技术语句的科技成果需求短语；Reconstruct the key technical sentences corresponding to the user's scientific and technological achievement demand information according to the preset phrase rules to obtain the scientific and technological achievement demand phrases of the corresponding key technical sentences;
将科技成果需求短语转换成相应的特征向量；Convert the scientific and technological achievement requirement phrases into corresponding feature vectors;
将各科技成果需求短语的特征向量作为相应科技成果需求信息的关键特征矩阵的元素，构建得到用户的科技成果需求信息的关键特征矩阵。The feature vectors of each scientific and technological achievement demand phrase are used as elements of the key feature matrix of the corresponding scientific and technological achievement demand information to construct the key feature matrix of the user's scientific and technological achievement demand information.
对于每一科技成果，通过以下方式构建得到科技成果的关键特征矩阵，执行：For each scientific and technological achievement, the key feature matrix of the scientific and technological achievement is constructed in the following way and executed:
提取科技成果中的技术关键词，将科技成果中包含任意技术关键词的语句作为包含技术关键词的科技成果语句；Extract technical keywords from scientific and technological achievements, and take sentences containing any technical keywords in scientific and technological achievements as sentences of scientific and technological achievements containing technical keywords;
对科技成果的科技成果语句分别按照预设短语规则进行重构，得到相应语句的科技成果短语；Reconstruct the scientific and technological achievement sentences of the scientific and technological achievements according to the preset phrase rules to obtain the scientific and technological achievement phrases of the corresponding sentences;
将科技成果短语转换成相应的特征向量，将各科技成果短语的特征向量作为科技成果的关键特征矩阵的元素，构建得到相应科技成果的关键特征矩阵。The scientific and technological achievement phrases are converted into corresponding feature vectors, and the feature vectors of each scientific and technological achievement phrase are used as elements of the key feature matrix of the scientific and technological achievements to construct the key feature matrix of the corresponding scientific and technological achievements.
本发明的有益效果在于：The beneficial effects of the present invention are:
本实施例提供的基于人工智能交互的科技成果推荐系统，通过人工智能交互模块对科技成果需求信息进行智能检索、扩展，之后将科技成果需求信息及科技成果联合需求信息转化成表征向量形式进行后续科技成果预测。由于对原始的科技成果需求信息进行了技术扩展丰富，此时预测得到的科技成果的范围更广、数据量更大。为避免预测得到的科技成果与原始需求信息的关联性较差，本实施还设置了输出控制模块，通过关键特征矩阵匹配的方式剔除掉匹配度较差的科技成果，从而得到最终推荐给用户的科技成果。因此，本实施例通过采用“宽进严出”的技术构思，前期对科技成果需求信息进行丰富，后期对筛选出的科技成果的匹配度进行把关，从而有效保证了科技成果推荐系统推荐的科技成果既能满足客户需求，又避免了因过于精准的科技成果推荐导致的信息茧房效应的问题。The scientific and technological achievements recommendation system based on artificial intelligence interaction provided in this embodiment intelligently retrieves and expands the scientific and technological achievements demand information through the artificial intelligence interaction module, and then converts the scientific and technological achievements demand information and the scientific and technological achievements joint demand information into the representation vector form for subsequent scientific and technological achievements prediction. Since the original scientific and technological achievements demand information has been technically expanded and enriched, the predicted scientific and technological achievements have a wider range and a larger amount of data. In order to avoid the poor correlation between the predicted scientific and technological achievements and the original demand information, this implementation also sets an output control module to eliminate the scientific and technological achievements with poor matching by matching the key feature matrix, so as to obtain the scientific and technological achievements finally recommended to the user. Therefore, this embodiment adopts the technical concept of "loose entry and strict exit", enriches the scientific and technological achievements demand information in the early stage, and checks the matching degree of the selected scientific and technological achievements in the later stage, thereby effectively ensuring that the scientific and technological achievements recommended by the scientific and technological achievements recommendation system can not only meet customer needs, but also avoid the problem of information cocoon effect caused by overly accurate scientific and technological achievements recommendation.
本实施例提供的基于人工智能交互的科技成果推荐系统的结构清晰、推荐结果准确，各个模块之间相互独立，可扩展性强，具有较好的推广价值和应用前景。The scientific and technological achievement recommendation system based on artificial intelligence interaction provided in this embodiment has a clear structure, accurate recommendation results, each module is independent of each other, has strong scalability, and has good promotion value and application prospects.
附图说明BRIEF DESCRIPTION OF THE DRAWINGS
为了更清楚地说明本发明实施例或现有技术中的技术方案，下面将对实施例或现有技术描述中所需要使用的附图作简单地介绍，显而易见地，对于本领域普通技术人员来讲，在不付出创造性劳动的前提下，还可以根据这些附图获得其他的附图。In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.
图1为本发明实施例提供的基于人工智能交互的科技成果推荐系统的结构示意图。FIG1 is a schematic diagram of the structure of a scientific and technological achievement recommendation system based on artificial intelligence interaction provided by an embodiment of the present invention.
具体实施方式DETAILED DESCRIPTION
下面结合附图来具体描述本发明的优选实施例，其中，附图构成本申请一部分，并与本发明的实施例一起用于阐释本发明的原理，并非用于限定本发明的范围。The preferred embodiments of the present invention are described in detail below in conjunction with the accompanying drawings, wherein the accompanying drawings constitute a part of this application and are used together with the embodiments of the present invention to illustrate the principles of the present invention, but are not used to limit the scope of the present invention.
本发明的一个具体实施例，公开了一种基于人工智能交互的科技成果推荐系统，科技成果推荐系统的结构示意图参考图1，包括以下组成部分：用户输入模块、人工智能交互模块、关联分析模块、向量构建模块、科技成果预测模块及输出控制模块；其中，用户输入模块，接收用户输入的科技成果需求信息；人工智能交互模块，利用人工智能交互系统对科技成果需求信息进行智能检索，获得科技成果联合需求信息；向量构建模块，基于科技成果需求信息及科技成果联合需求信息，构建科技成果需求表征向量；科技成果预测模块，将科技成果需求表征向量输入科技成果预测模型，由科技成果预测模型预测输出参考科技成果；输出控制模块，将科技成果需求信息的关键特征矩阵与参考科技成果的关键特征矩阵进行匹配判断，从参考科技成果中选出匹配判断通过的科技成果推荐给用户。A specific embodiment of the present invention discloses a scientific and technological achievement recommendation system based on artificial intelligence interaction. The structural schematic diagram of the scientific and technological achievement recommendation system refers to Figure 1, which includes the following components: a user input module, an artificial intelligence interaction module, an association analysis module, a vector construction module, a scientific and technological achievement prediction module and an output control module; wherein the user input module receives scientific and technological achievement demand information input by the user; the artificial intelligence interaction module uses the artificial intelligence interaction system to intelligently search the scientific and technological achievement demand information to obtain the scientific and technological achievement joint demand information; the vector construction module constructs a scientific and technological achievement demand representation vector based on the scientific and technological achievement demand information and the scientific and technological achievement joint demand information; the scientific and technological achievement prediction module inputs the scientific and technological achievement demand representation vector into the scientific and technological achievement prediction model, and the scientific and technological achievement prediction model predicts and outputs reference scientific and technological achievements; the output control module matches and judges the key feature matrix of the scientific and technological achievement demand information with the key feature matrix of the reference scientific and technological achievements, and selects the scientific and technological achievements that pass the matching judgment from the reference scientific and technological achievements and recommends them to the user.
具体实施过程中，为进一步提升科技成果推荐系统的使用便利性，所述科技成果推荐系统还包括可视化展示模块，用于将匹配判断通过的科技成果按照匹配度由高到低的顺序排序，进行可视化展示。During the specific implementation process, in order to further improve the ease of use of the scientific and technological achievement recommendation system, the scientific and technological achievement recommendation system also includes a visualization display module, which is used to sort the scientific and technological achievements that have passed the matching judgment in order from high to low matching degree for visualization.
下面，对科技成果推荐系统的主要模块组成做如下介绍：Below, the main modules of the scientific and technological achievements recommendation system are introduced as follows:
(1)用户输入模块(1) User input module
用户输入模块内置用户输入模板，用户依据用户输入模板输入科技成果需求信息；其中，所述用户输入模板包括但不限于以下内容：科技成果需求的技术领域、技术要点及技术关键词。通过设置交互输入模板，保证用户输入的内容更加全面、具体，便于推荐到更加符合用户需求的科技成果。The user input module has a built-in user input template, and the user inputs the scientific and technological achievement demand information according to the user input template; wherein, the user input template is not limited to the following content: the technical field, technical points and technical keywords of the scientific and technological achievement demand. By setting the interactive input template, it is ensured that the content input by the user is more comprehensive and specific, so as to facilitate the recommendation of scientific and technological achievements that better meet the user's needs.
(2)人工智能交互模块(2) Artificial Intelligence Interaction Module
人工智能交互模块内置人工智能交互模板，用于将用户输入的科技成果需求信息置入人工智能交互模板，形成人工智能交互语句；由人工智能交互系统以人工智能交互语句为输入，对科技成果需求信息进行智能检索，获得科技成果联合需求信息。The artificial intelligence interaction module has a built-in artificial intelligence interaction template, which is used to place the scientific and technological achievement demand information input by the user into the artificial intelligence interaction template to form an artificial intelligence interaction statement; the artificial intelligence interaction system uses the artificial intelligence interaction statement as input to perform intelligent retrieval of the scientific and technological achievement demand information to obtain joint scientific and technological achievement demand information.
示例性地，所述人工智能交互模板的样式设计为：检索“科技成果需求信息”的科技文献。需要说明的是，由于本实施例旨在实现科技成果推荐，因此，生成式人工智能交互系统的训练语料应包含大量的科技成果知识。常见的生成式人工智能交互系统包括ChatGPT、文心一言或chatGLM。Exemplarily, the style design of the artificial intelligence interaction template is: search for scientific and technological literature of "scientific and technological achievement demand information". It should be noted that since this embodiment aims to achieve scientific and technological achievement recommendation, the training corpus of the generative artificial intelligence interaction system should contain a large amount of scientific and technological achievement knowledge. Common generative artificial intelligence interaction systems include ChatGPT, Wenxinyiyan or chatGLM.
(3)向量构建模块(3) Vector building blocks
向量构建模块包括知识图谱构建单元和图谱表征单元；其中，The vector construction module includes a knowledge graph construction unit and a graph representation unit;
知识图谱构建单元，用于从科技成果需求信息及科技成果联合需求信息中提取出实体、实体属性及实体间的关系，并基于所述实体、实体属性及实体间的关系，构建知识图谱。具体实施过程中，可以使用图数据库构建知识图谱，如Neo4j等，来存储和管理图数据。The knowledge graph construction unit is used to extract entities, entity attributes and relationships between entities from the scientific and technological achievement demand information and the scientific and technological achievement joint demand information, and construct a knowledge graph based on the entities, entity attributes and relationships between entities. In the specific implementation process, a graph database, such as Neo4j, can be used to construct a knowledge graph to store and manage graph data.
图谱表征单元，使用图嵌入算法(如Node2Vec、GraphSAGE或TransE)对知识图谱进行图谱表征，构建科技成果需求表征向量。The graph representation unit uses a graph embedding algorithm (such as Node2Vec, GraphSAGE, or TransE) to represent the knowledge graph and construct a representation vector for scientific and technological achievement requirements.
(4)科技成果预测模块(4) Scientific and technological achievements prediction module
实际实施过程中，科技成果预测模块，将科技成果需求表征向量输入科技成果预测模型后，由科技成果预测模型根据科技成果库中各科技成果表征向量与科技成果需求表征向量的匹配程度(即，匹配度得分)，根据科技成果预测模型输出的匹配度得分对各科技成果进行排序。根据排序结果，选取分数最高的若干个科技成果作为参考科技成果进行预测输出。In the actual implementation process, the scientific and technological achievement prediction module inputs the scientific and technological achievement demand representation vector into the scientific and technological achievement prediction model. The scientific and technological achievement prediction model then sorts the scientific and technological achievements according to the matching degree (i.e., matching score) between the scientific and technological achievement representation vectors in the scientific and technological achievement database and the scientific and technological achievement demand representation vectors. According to the sorting results, several scientific and technological achievements with the highest scores are selected as reference scientific and technological achievements for prediction output.
科技成果预测模型通过以下方式训练得到：The scientific and technological achievement prediction model is trained in the following ways:
获取多组训练样本数据，每组样本数据包括用户的科技成果需求信息样本、科技成果样本、以及各科技成果样本与相应用户的科技成果需求信息样本的匹配度得分。需要说明的是，这里的科技成果样本既包含正科技成果样本(予以推荐，匹配度得分较高)，也包括负科技成果样本(不予推荐，匹配度得分较低)。获取样本数据中用户的科技成果需求信息样本、科技成果样本的表征向量；将用户科技成果需求信息样本的表征向量输入科技成果预测模型，并用以训练拟合相应科技成果样本的表征向量的匹配度得分。直至训练完成，得到训练好的科技成果预测模型。Obtain multiple groups of training sample data, each group of sample data includes the user's scientific and technological achievement demand information samples, scientific and technological achievement samples, and the matching scores of each scientific and technological achievement sample and the corresponding user's scientific and technological achievement demand information samples. It should be noted that the scientific and technological achievement samples here include both positive scientific and technological achievement samples (recommended, with a high matching score) and negative scientific and technological achievement samples (not recommended, with a low matching score). Obtain the user's scientific and technological achievement demand information samples and scientific and technological achievement sample representation vectors in the sample data; input the user's scientific and technological achievement demand information sample representation vector into the scientific and technological achievement prediction model, and use it to train and fit the matching score of the representation vector of the corresponding scientific and technological achievement sample. After the training is completed, a trained scientific and technological achievement prediction model is obtained.
需要说明的是，本实施例中的科技成果库是预先构建的，在将科技成果存入科技成果库的同时，可将科技成果的表征向量、技术领域、技术关键词、关键特征矩阵同时存入科技成果库。It should be noted that the scientific and technological achievements database in this embodiment is pre-built. When storing scientific and technological achievements in the scientific and technological achievements database, the characterization vectors, technical fields, technical keywords, and key feature matrices of the scientific and technological achievements can also be stored in the scientific and technological achievements database at the same time.
具体地，科技成果库作为科技成果详细信息的数据库，包括但不限于科技成果的概述、技术领域、成熟度、知识产权状况、获奖情况等。Specifically, the scientific and technological achievement database serves as a database of detailed information on scientific and technological achievements, including but not limited to an overview of the scientific and technological achievements, technical fields, maturity, intellectual property status, awards, etc.
在本实施例中，对于每一科技成果，具体执行以下操作，构建得到科技成果的表征向量：基于科技成果构建相应的知识图谱，之后通过图谱表征方式得到科技成果的表征向量。相关实现方式与用户的科技成果需求信息对应的表征向量的构建方式类似，此处不再赘述。In this embodiment, for each scientific and technological achievement, the following operations are specifically performed to construct a representation vector of the scientific and technological achievement: a corresponding knowledge graph is constructed based on the scientific and technological achievement, and then the representation vector of the scientific and technological achievement is obtained through graph representation. The relevant implementation method is similar to the construction method of the representation vector corresponding to the user's scientific and technological achievement demand information, and will not be repeated here.
(5)输出控制模块(5) Output control module
输出控制模块包括用户关键特征矩阵提取单元、匹配判断单元和筛选单元；其中，The output control module includes a user key feature matrix extraction unit, a matching judgment unit and a screening unit; wherein,
用户关键特征矩阵提取单元，用于从用户输入的科技成果需求信息中提取包含技术关键词的语句作为关键技术语句；并从关键技术语句中提取关键特征矩阵；A user key feature matrix extraction unit is used to extract sentences containing technical keywords as key technical sentences from the scientific and technological achievement demand information input by the user; and to extract a key feature matrix from the key technical sentences;
匹配判断单元，将科技成果需求信息的关键特征矩阵与参考科技成果的关键特征矩阵利用相似度原则进行匹配判断，得到各参考科技成果对应的匹配判断结果；The matching judgment unit matches and judges the key feature matrix of the scientific and technological achievement demand information with the key feature matrix of the reference scientific and technological achievements using the similarity principle to obtain the matching judgment results corresponding to each reference scientific and technological achievement;
筛选单元，根据预设匹配阈值，从参考科技成果中选出匹配判断结果超过匹配阈值的科技成果，并将其作为匹配判断通过的科技成果推荐给用户。The screening unit selects scientific and technological achievements whose matching judgment results exceed the matching threshold from the reference scientific and technological achievements according to the preset matching threshold, and recommends them to the user as scientific and technological achievements that pass the matching judgment.
具体实施过程中，从关键技术语句中提取关键特征矩阵，执行：In the specific implementation process, the key feature matrix is extracted from the key technical statements and executed:
对用户的科技成果需求信息对应的关键技术语句按照预设短语规则进行重构，得到相应关键技术语句的科技成果需求短语；Reconstruct the key technical sentences corresponding to the user's scientific and technological achievement demand information according to the preset phrase rules to obtain the scientific and technological achievement demand phrases of the corresponding key technical sentences;
将科技成果需求短语转换成相应的特征向量；Convert the scientific and technological achievement requirement phrases into corresponding feature vectors;
将各科技成果需求短语的特征向量作为相应科技成果需求信息的关键特征矩阵的元素，构建得到用户的科技成果需求信息的关键特征矩阵。The feature vectors of each scientific and technological achievement demand phrase are used as elements of the key feature matrix of the corresponding scientific and technological achievement demand information to construct the key feature matrix of the user's scientific and technological achievement demand information.
具体实施过程中，对于每一科技成果，通过以下方式构建得到科技成果的关键特征矩阵，执行：During the specific implementation process, for each scientific and technological achievement, the key feature matrix of the scientific and technological achievement is constructed and executed in the following way:
提取科技成果中的技术关键词，将科技成果中包含任意技术关键词的语句作为包含技术关键词的科技成果语句；Extract technical keywords from scientific and technological achievements, and take sentences containing any technical keywords in scientific and technological achievements as sentences of scientific and technological achievements containing technical keywords;
对科技成果的科技成果语句分别按照预设短语规则进行重构，得到相应语句的科技成果短语；Reconstruct the scientific and technological achievement sentences of the scientific and technological achievements according to the preset phrase rules to obtain the scientific and technological achievement phrases of the corresponding sentences;
将科技成果短语转换成相应的特征向量，将各科技成果短语的特征向量作为科技成果的关键特征矩阵的元素，构建得到相应科技成果的关键特征矩阵。The scientific and technological achievement phrases are converted into corresponding feature vectors, and the feature vectors of each scientific and technological achievement phrase are used as elements of the key feature matrix of the scientific and technological achievements to construct the key feature matrix of the corresponding scientific and technological achievements.
综上所述，本实施例提供的基于人工智能交互的科技成果推荐系统，通过人工智能交互模块对科技成果需求信息进行智能检索、扩展，之后将科技成果需求信息及科技成果联合需求信息转化成表征向量形式进行后续科技成果预测。由于对原始的科技成果需求信息进行了技术扩展丰富，此时预测得到的科技成果的范围更广、数据量更大。为避免预测得到的科技成果与原始需求信息的关联性较差，本实施还设置了输出控制模块，通过关键特征矩阵匹配的方式剔除掉匹配度较差的科技成果，从而得到最终推荐给用户的科技成果。因此，本实施例通过采用“宽进严出”的技术构思，前期对科技成果需求信息进行丰富，后期对筛选出的科技成果的匹配度进行把关，从而有效保证了科技成果推荐系统推荐的科技成果既能满足客户需求，又避免了因过于精准的科技成果推荐导致的信息茧房效应的问题。In summary, the scientific and technological achievements recommendation system based on artificial intelligence interaction provided in this embodiment intelligently retrieves and expands the scientific and technological achievements demand information through the artificial intelligence interaction module, and then converts the scientific and technological achievements demand information and the scientific and technological achievements joint demand information into the representation vector form for subsequent scientific and technological achievements prediction. Since the original scientific and technological achievements demand information is technically expanded and enriched, the predicted scientific and technological achievements at this time have a wider range and a larger amount of data. In order to avoid the poor correlation between the predicted scientific and technological achievements and the original demand information, this implementation also sets an output control module, and eliminates the scientific and technological achievements with poor matching by matching the key feature matrix, so as to obtain the scientific and technological achievements finally recommended to the user. Therefore, this embodiment adopts the technical concept of "loose entry and strict exit", enriches the scientific and technological achievements demand information in the early stage, and checks the matching degree of the selected scientific and technological achievements in the later stage, thereby effectively ensuring that the scientific and technological achievements recommended by the scientific and technological achievements recommendation system can not only meet customer needs, but also avoid the problem of information cocoon effect caused by overly accurate scientific and technological achievements recommendation.
本实施例提供的基于人工智能交互的科技成果推荐系统的结构清晰、推荐结果准确，各个模块之间相互独立，可扩展性强，具有较好的推广价值和应用前景。The scientific and technological achievement recommendation system based on artificial intelligence interaction provided in this embodiment has a clear structure, accurate recommendation results, each module is independent of each other, has strong scalability, and has good promotion value and application prospects.
对于本领域技术人员而言，可以理解在不脱离本发明的原理和精神的情况下可以对这些实施例进行多种变化、修改、替换和变型。因此，从任意一处来说，都应将实施例看作是指导性的，而且是非限制性的，本发明的范围由所附权利要求而不是上述说明限定，本发明的较佳实施例而已，并不用以限制本发明，凡在本发明的精神和原则之内，所有的任何修改、等同替换、改进等，均应包含在本发明的保护范围之内。For those skilled in the art, it is understood that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the present invention. Therefore, from any point of view, the embodiments should be regarded as instructive and non-restrictive. The scope of the present invention is limited by the attached claims rather than the above description. It is only a preferred embodiment of the present invention and is not intended to limit the present invention. All modifications, equivalent substitutions, improvements, etc. within the spirit and principles of the present invention should be included in the scope of protection of the present invention.
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410694847.8ACN118467792A (en) | 2024-05-31 | 2024-05-31 | A scientific and technological achievement recommendation system based on artificial intelligence interaction |
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410694847.8ACN118467792A (en) | 2024-05-31 | 2024-05-31 | A scientific and technological achievement recommendation system based on artificial intelligence interaction |
| Publication Number | Publication Date |
|---|---|
| CN118467792Atrue CN118467792A (en) | 2024-08-09 |
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410694847.8APendingCN118467792A (en) | 2024-05-31 | 2024-05-31 | A scientific and technological achievement recommendation system based on artificial intelligence interaction |
| Country | Link |
|---|---|
| CN (1) | CN118467792A (en) |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105653706A (en)* | 2015-12-31 | 2016-06-08 | 北京理工大学 | Multilayer quotation recommendation method based on literature content mapping knowledge domain |
| US20190122145A1 (en)* | 2017-10-23 | 2019-04-25 | Baidu Online Network Technology (Beijing) Co., Ltd. | Method, apparatus and device for extracting information |
| CN114218499A (en)* | 2022-02-22 | 2022-03-22 | 腾讯科技（深圳）有限公司 | Resource recommendation method and device, computer equipment and storage medium |
| CN117743565A (en)* | 2024-01-31 | 2024-03-22 | 广东省华南技术转移中心有限公司 | Scientific and technological achievement recommendation method, device and storage medium |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105653706A (en)* | 2015-12-31 | 2016-06-08 | 北京理工大学 | Multilayer quotation recommendation method based on literature content mapping knowledge domain |
| US20190122145A1 (en)* | 2017-10-23 | 2019-04-25 | Baidu Online Network Technology (Beijing) Co., Ltd. | Method, apparatus and device for extracting information |
| CN114218499A (en)* | 2022-02-22 | 2022-03-22 | 腾讯科技（深圳）有限公司 | Resource recommendation method and device, computer equipment and storage medium |
| CN117743565A (en)* | 2024-01-31 | 2024-03-22 | 广东省华南技术转移中心有限公司 | Scientific and technological achievement recommendation method, device and storage medium |
| Publication | Publication Date | Title |
|---|---|---|
| CN110489538B (en) | Statement response method and device based on artificial intelligence and electronic equipment | |
| Wang et al. | Patentformer: a novel method to automate the generation of patent applications | |
| CN118964589A (en) | A method for constructing an intelligent search engine based on a large language model | |
| Ding et al. | A knowledge-enriched and span-based network for joint entity and relation extraction | |
| CN117312493A (en) | Multi-strategy knowledge extraction system | |
| Li et al. | Biomedical named entity recognition based on the two channels and sentence-level reading control conditioned LSTM-CRF | |
| Zheng et al. | Chinese medical named entity recognition using crf-mt-adapt and ner-mrc | |
| CN112148857B (en) | A system and method for automatically generating official documents | |
| CN112287656A (en) | Text comparison method, device, equipment and storage medium | |
| CN119047468B (en) | Intelligent creation method based on NLP intention recognition | |
| CN113127627B (en) | Poetry Recommendation Method Based on LDA Topic Model and Poetry Knowledge Graph | |
| CN118917294B (en) | Intelligent writing method and system based on multi-role intelligent agent | |
| CN114328820B (en) | Information search method and related equipment | |
| CN118350368B (en) | Multi-document select and edit method of large language model based on NLP technology | |
| CN118332392B (en) | Multi-task mental health identification method and system integrating prior knowledge and expert network | |
| CN112559711B (en) | Synonymous text prompting method and device and electronic equipment | |
| CN118606429A (en) | A method for accurate recommendation of scientific and technological achievements | |
| CN119128125A (en) | Personalized document recommendation method and system based on knowledge graph | |
| CN118709654A (en) | A method for generating event summary text based on graph neural network based on topic enhancement | |
| Kan et al. | A Unified Generative Framework based on Prompt Learning for Various Information Extraction Tasks | |
| CN118467792A (en) | A scientific and technological achievement recommendation system based on artificial intelligence interaction | |
| CN112861956A (en) | Water pollution model construction method based on data analysis | |
| CN118484506A (en) | Business identification system based on semantic analysis | |
| CN116011441B (en) | Keyword extraction method and system based on pre-training model and automatic receptive field | |
| CN116775855A (en) | Automatic TextRank Chinese abstract generation method based on Bi-LSTM |
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |