Product
Eureka
For R&D, Eureka makes reading and utilizing patents & technical documents easy.
Eureka AIR
Designed for self-driven R&D workflows. Generate viable solutions, solve complex R&D challenges, empower your innovation with AI.
Eureka Materials
Designed for material experts only. Revolutionize your material R&D, from search, analyze, to developing new materials.
Feature
TechResearch
Generate reliable direction feasibility study reports for your R&D in just a few steps.
TechSeek
Discover and master advanced knowledge NOW. Basics, ideas, possibilities, all at once.
TechMind
As an expert in R&D Theories, TechMind can generates customized viable solutions instantly.
TechRisk
Analyze your overall solution with one click, know your potential R&D risks in advance.
TechMonitor
Get weekly tech updates, stay abreast of the latest tech innovations and key insights.
Patents
Literature
Hiro is an intelligent assistant for R&D personnel, combined with Patent DNA, to facilitate innovative research.
38 results about "Traditional knowledge" patented technology
Efficacy Topic
Property
Owner
Technical Advancement
Application Domain
Technology Topic
Technology Field Word
Patent Country/Region
Patent Type
Patent Status
Application Year
Inventor
Traditional knowledge, indigenous knowledge and local knowledge generally refer to knowledge systems embedded in the cultural traditions of regional, indigenous, or local communities. Traditional knowledge includes types of knowledge about traditional technologies of subsistence (e.g. tools and techniques for hunting or agriculture), midwifery, ethnobotany and ecological knowledge, traditional medicine, celestial navigation, ethnoastronomy, climate, and others. These kinds of knowledge, crucial for subsistence and survival, are generally based on accumulations of empirical observation and on interaction with the environment.
Dynamic legal event map construction method oriented to legal field
ActiveCN109977237ANatural language data processingNeural architecturesTraditional knowledgeKnowledge extraction
The invention discloses a dynamic legal event map construction method oriented to the legal field, and a dynamic legal event map oriented to the legal field is constructed by utilizing a natural language processing technology. In the prior art, the dynamic legal event map construction method does not appear in the legal field, compared with the prior art, the minimum unit of the dynamic legal event map of the present invention is a legal event rather than an entity, and the constructed map is composed of a relationship between the legal event and the legal event. Compared with the traditionalevent extraction, the knowledge in the constructed dynamic legal event map is dynamic, computable and reasonable instead of a simple extraction event. According to the method, the dynamic legal eventmap oriented to the legal field is constructed, the map is constructed by combining a traditional knowledge map knowledge extraction technology and an event mining technology, and then the legal eventmap metering is carried out by modeling the legal event elements in the legal event map and the relations among the legal events respectively.
Owner:南京擎盾信息科技有限公司
A knowledge graph embedding method based on a diverse graph attention mechanism
ActiveCN109902183AAchieve acquisitionReasonable and Interpretable EmbeddingsSemantic tool creationSelf attentionTraditional knowledge
The invention provides a knowledge graph embedding method based on a diverse graph attention mechanism, and the method comprises the following steps: 1, carrying out the preprocessing of knowledge graph data, and enabling a traditional knowledge graph to be preprocessed into structured data according to model requirements; 2, establishing an entity attention mechanism, wherein the entity attentionmechanism establishing mode is that the vector representation of an entity is learned on an adjacent graph by using n attention heads and self-attention through using a diverse graph attention mechanism; Step 3, establishing a relation attention mechanism in which the vector representation of the entity on the relation graph is learned by using n attention heads and self-attention by using a diverse graph attention mechanism; Step 4, carrying out modeling of multi-relation data in the relation graph; And 5, carrying out model training to obtain vectorized representations of all relationshipsin the knowledge graph.
Owner:BEIHANG UNIV
Iterative entity alignment model
The invention provides an entity alignment method and device. The method and device aim to solve the problem that traditional knowledge alignment method usually requires cumbersome manual labor or elaborate manual construction. The method comprises the steps of acquiring vector representations of entities in a first knowledge graph and vector representations of the entities in a second knowledge graph separately, binding vector representations, with the same meanings, of the entities in the first knowledge graph and the second knowledge graph according to an input alignment seed entity pair set so as to acquire a third knowledge graph, performing iterative calculation of the similarity among the vector representations of the entities according to the third knowledge graph, the vector representations of the entities in the first knowledge graph and the vector representations of the entities in the second knowledge graph, in the process of iterative calculation, if the distance between a certain pair of vector representations of the entities is less than a threshold, adding the entity corresponding to the pair of vector representations to the third knowledge graph until the number of entities in the third knowledge graph is no longer increased. The entity alignment method and device are good in practicality.
Owner:TSINGHUA UNIV
Knowledge representation learning method fusing relationship path, type and entity description information
ActiveCN108763376AClear semanticsSemantic characterizationSpecial data processing applicationsPositive sampleTraditional knowledge
The invention discloses a knowledge representation learning method fusing relationship path, type and entity description information. The method comprises the following steps: (1) preprocessing a knowledge base to extract structured information, relationship path information, type information and entity description information; (2) selecting a positive sample from the knowledge base, constructinga negative sample, calculating the loss of the structured information, the relationship path information, the type information and the entity description information corresponding to the positive andnegative samples; (3) updating parameters, a practice vector and a relationship vector of a model according to the loss; and (4) repeating the step (1) to the step (3) until the number of iterations reaches the preset maximum number of iterations, and outputting the vectors of the entity and relationship to realize the learning of the knowledge representation. According to the method, on the basisof using a relationship path, modelling is performed on the knowledge base by fusing multi-source information, so that the problem that the only the structured information is used in traditional knowledge representation learning method and multiple additional information is not adopted is solved.
Owner:ZHEJIANG UNIV
Feature tensor-based Chinese knowledge graph representation learning method
ActiveCN111160564AHigh precisionFast convergenceMachine learningSpecial data processing applicationsAlgorithmTheoretical computer science
The invention provides a feature tensor-based Chinese knowledge graph representation learning method. The method comprises the following steps: carrying out data preparation, establishing a data structure, constructing an entity feature vector matrix, definiing relation vectors and distance formulas of marked triples, obtaining a training set, training a knowledge graph representation learning model, updating model parameters and iterative training, and carrying out relationship prediction on the unmarked triad by using the model, and carrying out iterative training again until a new unmarkedtriad cannot be learned. According to the method, feature tensors are formed by using Chinese pinyin, character information, word information and description information and are converted into featurevectors, so that a method for randomly initializing entity vectors in traditional knowledge representation learning is replaced, and Chinese features are fully utilized. Besides, a double-layer iteration mode is adopted to supplement the training corpus, so that the relation matrix can be continuously corrected, and the precision and convergence speed of the knowledge graph representation learning model are improved.
Owner:UNIV OF ELECTRONICS SCI & TECH OF CHINA
Theme map generation method, device and equipment suitable for text analysis or data mining and computer storage medium
ActiveCN109902302AMake up depthMake up for the lack of visualizationSpecial data processing applicationsSemantic tool creationTopic MapsNODAL
The invention relates to the technical field of computers, and discloses a topic map generation method and device suitable for text analysis or data mining, equipment and a computer storage medium. Inthis way, the invention provides a new method for enabling the application of the probability topic model LDA to reach a brand new level, so each theme can become a node in a semantic network; the complex semantic association among the topics is reflected more deep; the knowledge graph and the traditional knowledge graph are mutually corresponding and complemented; the theme map having an independent application value is achieved; Therefore, the efficiency and accuracy of traditional search and recommendation can be effectively improved, the defects of depth and visualization of traditional data mining or text analysis can be overcome, potential semantic association difficult to find by a traditional method can be mined out, real scientific discovery is achieved, and particularly, the method has a huge potential value in data mining.
Owner:郑敏杰
Knowledge service innovation method driven by polymorphic knowledge graph
PendingCN112199515AProblem solving problem solving process relies on experienceKnowledge representationSpecial data processing applicationsPersonalizationTraditional knowledge
The invention discloses a knowledge service innovation party driven by a polymorphic knowledge graph. The knowledge service innovation party comprises the following steps: 1) constructing a polymorphic domain knowledge graph; constructing a knowledge model, extracting and storing polymorphic knowledge, and providing a candidate scheme set for knowledge services; and 2) with an enterprise production problem task as a drive, modeling the process of solving the engineering problem of the knowledge service innovation method template from three aspects of a problem solving activity flow, a knowledge flow and a process decision flow by means of a scientific workflow model so as to support and guide the improvement activity of an enterprise in a template application layer. According to the method, a theoretical basis is provided for application of a knowledge service innovation method, and the problems of high subjectivity and high experience dependence in traditional knowledge service application are solved to a certain extent; the multi-knowledge service flexible template can be configured through the method, and a systematic innovative solving thought is provided for enterprises for complex and changeable personalized problems.
Owner:XI AN JIAOTONG UNIV
Visual question-answering method and system based on external knowledge aggregation
ActiveCN111475656AImprove applicabilityExcellent Visual Question Answering AccuracyCharacter and pattern recognitionEnergy efficient computingTheoretical computer scienceTraditional knowledge
The invention discloses a visual question-answering method and system based on external knowledge aggregation. The method comprises the steps that firstly, an external knowledge graph subgraph relatedto a scene is extracted, then knowledge aggregation is conducted on the knowledge graph subgraph to obtain knowledge entity representation, and finally the entity representation and a traditional knowledge question-answering system are organically fused to finally obtain a question answer. According to the method, the external knowledge graph is introduced into a traditional visual question-answering system, the method can be applied to traditional visual questions and visual questions needing external knowledge, extra strong supervision information is not needed in the model training process, the method has high applicability, and better visual question accuracy can be obtained on multiple benchmark data sets.
Owner:TSINGHUA UNIV
Knowledge distillation vertical field detection method based on similarity maintenance
ActiveCN113673254AImprove performanceReduce overheadSemantic analysisNeural architecturesTraditional knowledgeEngineering
The invention discloses a knowledge distillation standing field detection method based on similarity maintenance, and belongs to the technical field of text emotion detection. The method aims to solve the problems that an existing vertical field detection method is large in computing resource and computing time expenditure and the recognition accuracy needs to be improved in the existing method. According to the method, the Text-CNN network model is utilized to carry out vertical field detection on a to-be-detected vertical field text; a Text-CNN network is determined based on a knowledge distillation mode, namely, a BERT model is used as a teacher model, and implicit knowledge learned from the teacher model is introduced into a student model text; and a loss function based on similarity maintenance is provided on the basis of a traditional knowledge distillation mode, so that the performance of the student model obtained on the basis of the knowledge distillation mode is further improved. The method is mainly used for text standing field detection.
Owner:NORTHEAST FORESTRY UNIVERSITY
Named entity identification method based on knowledge graph
PendingCN111241840AImprove accuracyOptimize the representation of extracted entitiesNatural language data processingInference methodsPersonalizationNamed-entity recognition
The invention discloses a named entity recognition method based on a knowledge graph; the method carries out the reasoning of a relation indicator r and a tail entity t based on entity relation triples (h, r, t) in an existing knowledge graph. On the basis, a new model is provided to be used for representation learning among triple entity relationships extracted from an open domain knowledge base,so that relationship mining and knowledge reasoning in a knowledge graph are realized, and the effects of optimizing association search and personalized recommendation effects are achieved. Comparedwith a traditional knowledge base, the open domain knowledge base uses the relation indicator words to replace the relation types, entities are richer, and the granularity is finer and smoother.
Owner:中科曙光(南京)计算技术有限公司
Knowledge tracking method based on test question heterogeneous graph representation and learner embedding
ActiveCN113344053AScientific predictionComprehensive forecastCharacter and pattern recognitionNatural language data processingTracking modelTraditional knowledge
The invention relates to the field of education big data mining, graph neural network and learner behavior modeling, and provides a knowledge tracking method based on test question heterogeneous graph representation and learner embedding, the method adopts a heterogeneous graph neural network technology in the deep learning field to represent multi-dimensional features of test questions, and meanwhile, complex learner characteristics are modeled in combination with a project reaction theory, and ability characteristics of learners are captured by adopting a clustering mode and the like; finally, the test question and learner mixed features are fused to a traditional knowledge tracking model, and knowledge tracking and learner performance prediction for different learner groups are achieved. According to the invention, the learning condition of the learner can be scientifically and comprehensively predicted, and the purpose of assisting the teacher in precise teaching is achieved.
Owner:HUAZHONG NORMAL UNIV
Cause map with time node
PendingCN110489564AImprove applicabilityEasy to combSpecial data processing applicationsSemantic tool creationExtensibilityTraditional knowledge
The invention relates to a cause map with time nodes, which comprises a time axis composed of a plurality of time nodes, and event nodes respectively and correspondingly connected with the plurality of time nodes on the time axis, and the event nodes are connected with one another according to corresponding constraint conditions. The cause takes time as a node in the map rather than as an attribute of an event. Compared with a traditional knowledge graph or event graph, the cause map has obvious expandability and the calculation performance is remarkably improved, and the application effect isgood.
Owner:贳巽(北京)国际商业数据技术股份公司
Deep embedded knowledge tracking method based on exercise difficulty and student ability
PendingCN113033808AMuch attentionNeural architecturesKnowledge representationDeep knowledgeTracking model
The invention belongs to the field of knowledge tracking, and discloses a deep embedded knowledge tracking method based on exercise difficulty and student ability, which comprises the following steps: S1, constructing a deep knowledge tracking model: integrating exercise difficulty features and student ability features into the deep knowledge tracking model; and S2, predicting student performance: fusing an attention mechanism, and predicting the student performance by using the deep knowledge tracking model constructed in the step 1. Compared with a traditional knowledge tracking method, the method has the advantages that the influence of different exercise difficulty and student ability on the answering condition of the student is considered, the model is enabled to focus more attention on answering records with similar exercise difficulty and student ability through an attention mechanism, and the prediction accuracy is improved. According to the method, the accuracy of knowledge tracking is improved, the method can be used in subsequent exercise recommendation application, and the application effect is improved.
Owner:NORTHWEST UNIV(CN)
Disinfection antibacterial agent
ActiveCN102406858APromote divisionRapid recovery splitAntibacterial agentsAntisepticsBiotechnologyCymbopogon distans
The invention provides a disinfection antibacterial agent, which is characterized by being prepared from the following components in parts by weight: 20-40 g of safflower, 10-20 g of cymbopogon distans, 20-30 g of lithospermum, 20-30 g of acanthopanax, 10-20 g of khasi nightshade fruit and 1000 mL of alcohol. On the basis of clinical practice and summarization and analysis of traditional knowledge of traditional Chinese medicinal plants, the pure natural medicinal plant extract is successfully researched by creatively selecting safflower, cymbopogon distans and other traditional Chinese medicinal plants as main raw materials and combining modern natural product extraction technique and bioengineering fermentation technique; and the disinfection antibacterial agent has good treatment effect or rehabilitation effect in antibacterial disinfection, wound healing, skin ulceration repairing, pain alleviation, resistance against tumors and other diseases or other fields.
Owner:向飞丹州
Knowledge graph question-answering system for professional field
ActiveCN113220861AAchieve retrievalImplementing Question Answering for Cross-Modal Knowledge RetrievalNatural language data processingSpecial data processing applicationsSemantic matchingTraditional knowledge
The invention discloses a knowledge graph question-answering system for the professional field. The system comprises a knowledge extraction module and an answer generation module, the processing process mainly comprises the following steps of extracting cross-modal collaborative professional field knowledge: performing knowledge extraction by utilizing text modal collaborative image and video modal data so as to solve the problem that a traditional knowledge extraction method is difficult to be directly applied to a professional field due to multi-source isomerism and multi-modality; questions and answers of multi-source knowledge are fused, and multi-source candidate answers are obtained through semantic matching and answer generation by utilizing professional domain knowledge information; and semantic fusion of the multi-source candidate answers, and generation of a final answer based on answer confidence reordering and answer key point reorganization of the multi-source knowledge. The knowledge graph question-answering system for the professional field solves the knowledge question-answering of multi-source heterogeneous and multi-modal data in the professional field, and is suitable for the multi-modal knowledge question-answering in other fields.
Owner:NO 54 INST OF CHINA ELECTRONICS SCI & TECH GRP
A dynamic legal event graph construction method for the legal field
ActiveCN109977237BNatural language data processingNeural architecturesTraditional knowledgeKnowledge extraction
Owner:南京擎盾信息科技有限公司
Call center customer service system knowledge base system method
InactiveCN107451944AIncrease speedImprove accuracyData processing applicationsSpecial data processing applicationsKnowledge-based systemsTraditional knowledge
The invention relates to a call center customer service system knowledge base system method. The method comprises the steps that a customer service executive understanding model is established; a two-level penetration system is established; a customer problem system, a business scenario system and a service situation system are integrated; and a speech flow chart model is established. According to the scheme of the invention, the disadvantages of a traditional knowledge base system are solved; the work difficulty of a customer service executive is reduced; and the work efficiency is improved.
Owner:国家电网有限公司客户服务中心南方分中心
Robot knowledge base construction system and method based on synonymous multi-context analysis
The invention provides a robot knowledge base construction system and method based on synonymous multi-context analysis, which realize the scheme from seven aspects of traditional knowledge base arrangement, question bank word segmentation, response bank word segmentation, synonymous expansion, context analysis, knowledge base indexing and new knowledge base construction. The method comprises: firstly, arranging an original knowledge base; performing word segmentation processing on the questioning sentences and the answering sentences in the knowledge base, and extracting keywords; performingsimilarity calculation and synonymy expansion on the keywords; determining adaptation scores of the keywords in the statement at the same time through context analysis, and constructing a multi-way tree index keyword at the same time; and finally, constructing a new knowledge base, and flexibly and randomly replacing synonyms in the response words by realizing similarity calculation of the synonyms, thereby realizing construction of the knowledge base with unchanged response theme content and diversified response modes.
Owner:STATE GRID CORP OF CHINA +2
A knowledge map management method and system based on semantic space mapping
InactiveCN104035917BIncreased processing flexibilityImprove accuracySpecial data processing applicationsSemantic vectorGraph spectra
Owner:FUDAN UNIV
Method for providing a platform for building a dynamic social knowledgebase
The present invention provides a method and system for providing a platform for building a dynamic social knowledgebase. The method includes importing a traditional knowledgebase from an entity at regular intervals, receiving login credentials from a user via a web portal or mobile application. Retrieving social knowledgebase from a social network associated with the user, providing recommendations to the user based on articles that are more clickable in the traditional knowledgebase and the social knowledgebase, sending a notification to the user based on the latest happenings in the web portal or mobile application, monitoring a list of top contributors and influencers of the social knowledgebase in the social network associated with the user, and sending the list of top contributors and influencers to a backend system for analysis & use.
Owner:CRMIT SOLUTIONS PVT LTD
A Knowledge Extraction Method Based on Online Encyclopedia Linked Entities
InactiveCN103902649BEasy to handleSpecial data processing applicationsText database clustering/classificationTraditional knowledgeKnowledge extraction
Owner:FUDAN UNIV
Kung fu noodles
PendingCN105995497AThe production method is completeNew production methodFood scienceTraditional knowledgeSugar
The invention provides Kung fu noodles with tough and chewy taste and unique workmanship. The noodles are prepared from 1000 g of flour, 100 g of eggs, 15 g of salt, 50 g of white sugar and 440 g of water. The manufacturing method of the front part of India flying cakes is connected and extended, and a manufacturing method for the complete and brand new noodle variety is created. The noodles are tougher and chewier compared with most noodles; compared with traditional noodles, the noodles are stretched for many times, are light and thin and have the higher ornamental value.
Owner:蒋其林
Construction method of power plant equipment knowledge base
InactiveCN110188087ARealize accumulationAchieving Fusion MigrationResourcesSpecial data processing applicationsPower stationProcess Measures
The invention relates to the technical field of power generation equipment operation, in particular to a construction method of a power plant equipment knowledge base, which comprises the following steps of firstly integrating the power plant side equipment asset data into a group database, and then obtaining the equipment asset data through an interface program; calculating an equipment operationindex in real time by using the original operation parameters of the power generation equipment, and storing the calculated index; forming the indexes and the parameters into the criteria according to the experiences of the field professionals, and arranging the criteria into a format in the system for implementation; combining and storing the equipment alarm, fault information, fault causes, processing measures and criteria in a rule form to form a rule base. Compared with a traditional knowledge base, the knowledge base of the scheme has more intelligent inheritance and practicability, is not limited to a traditional knowledge obtaining mode any more, and realizes a new knowledge inheritance mode, namely human-computer-human, and provides a basis for the power plant operation personnelto analyze the current equipment operation state.
Owner:上海长庚信息技术股份有限公司
Pearl four-season health-care mobile phone card and preparation method thereof
PendingCN113940975AImprove actual functionsImprove the function of the spleen and stomachSenses disorderNervous disorderHuman bodyEfficacy
According to the invention, traditional knowledge of traditional Chinese medicines is inherited and developed, special effects of pearl, perilla stems and bighead atractylodes rhizomes are scientifically utilized, and the health-care mobile phone card has different warm and cool properties to adapt to different seasons of spring, summer, autumn and winter; the mobile phone card can volatilize effective components of medicines through mobile phones, human bodies and natural temperature, effective components of the medicines slightly improve the transportation and transformation functions of the spleen and the stomach, regulate qi and harmonize the stomach, nourish yin and improve eyesight, calm the liver and suppress yang, soothe the nerves and arrest convulsion, slowly, lastingly and instantly make up the loss of body fluid and Yang qi caused by long-time mobile phone playing, and improve the function of the human body, so that the symptoms such as spleen qi weakness, insufficient cerebral blood supply, dry eyes, discomfort and restlessness are relieved; and the problems that the prior art is unsafe, imperfect, inconvenient, unbalanced, incapable of being widely applied and the like are solved. The scheme is scientific, simple, convenient, economical, instant and long-lasting, and positive social value can be certainly generated due to wide public requirements, safe and unique performance and remarkable health-care value.
Owner:湖北中庆生物科技有限公司
Knowledge extraction method of case-related property knowledge base
PendingCN113918682AImprove accuracyData processing applicationsNatural language data processingTemplate matchingTemplate based
The invention discloses a knowledge extraction method of a case-related property knowledge base. The method comprises the following steps: S10, constructing three knowledge extraction templates with different dimensions; S20, extracting a quintuple relationship set from the law and regulation text by utilizing the knowledge extraction templates; and S30, calculating confidence coefficients of extraction results of the three knowledge extraction templates by applying fuzzy logic, and enabling the three knowledge extraction templates to compete with one another to obtain a final extraction result. The three knowledge extraction templates are designed in different dimensions, the quintuple relationship is automatically extracted from law and regulation texts, and the confidence coefficient of the extraction result of each template is calculated by applying fuzzy logic, so that the three templates compete with one another, and compared with traditional knowledge extraction based on template matching, the accuracy of the extraction result is further improved.
Owner:SICHUAN UNIV
Knowledge system for HPV diagnosis, diagnosis opinion system and storage medium
PendingCN113889260AAchieve replicationMedical automated diagnosisText database queryingTraditional knowledgeKnowledge combination
The invention provides a knowledge system for HPV diagnosis. The knowledge system comprises a first knowledge base, a second knowledge base and a third knowledge base, wherein the first knowledge base is used for storing conclusion knowledge Ai; the second knowledge base is used for storing phenomenon type knowledge Bj, and the phenomenon type knowledge Bj can be arranged and combined into a plurality of phenomenon type knowledge combinations Ck; and the third knowledge base is used for storing a mapping relation Fk from the phenomenon type knowledge combination Ck to the conclusion type knowledge Ai. The knowledge system has the beneficial effects that the knowledge system abandons an information storage electronization mode of a traditional knowledge base. A classification mode of medical diagnosis and treatment guide information is provided, under the classification mode, a judgment mode which is originally formed by independent thinking of experts and field participation between phenomenon type knowledge and conclusion type knowledge is presented as a diagnosis and treatment mapping relation by computer logic, the links in which experts participate in HPV diagnosis and treatment can be copied and run efficiently through a computer program, and copying of top-level experts is achieved to a certain extent.
Owner:杨惠 +1
Disinfection antibacterial agent
ActiveCN102406858BPromote divisionRapid recovery splitAntibacterial agentsAntisepticsBiotechnologyCymbopogon distans
Owner:向飞丹州
Deep knowledge-project joint tracking-based learner portraying method and application thereof
PendingCN113989080AOvercoming inexplicable propertiesOvercome the disadvantage that it can only be applied to a fixed set of itemsData processing applicationsNeural architecturesDeep knowledgeConceptual graph
The invention discloses a deep knowledge-project joint tracking-based learner portraying method and application thereof. The method comprises the following steps: receiving learning behavior data from a learner, and extracting learner information and item information in the learning behavior data, including a learner identity identifier, an item identifier, an item state, a learner answer result and an original learner portrait; and enabling the extracted information to pass through a deep knowledge-project joint tracking network, and updating the learner portrait of the learner in real time. According to the invention, an integrated framework of knowledge tracking and a conceptual graph is constructed, and inaccurate user portraits caused by local missing of learning data are avoided; besides, the deep knowledge-project joint tracking in the invention overcomes the uninterpretable property of the hidden state of the traditional knowledge tracking algorithm, and integrates the course knowledge point inline state in the model, thereby further improving the accuracy of generating the learner portrait; in addition, the invention can be applied to a learning platform application scene with dynamically increased items.
Owner:ZHEJIANG UNIV
A data self-distillation method based on batch processing knowledge ensemble (BAKE)
PendingCN113869434AImplement canonical storageLow costCharacter and pattern recognitionNeural architecturesBatch processingNetwork architecture
The invention discloses the technical field of target detection, and particularly relates to a data self-distillation method based on batch processing knowledge ensemble (BAKE), which comprises the following steps: S1, giving a small batch of samples and a network in training, and storing similar samples; s2, avoiding self-cognition enhancement, and normalizing the affinity matrix A; s3, predicting samples in one batch, and then obtaining and forming a soft learning target; s4, realizing complete integration of knowledge. Compared with an existing knowledge integration method, the problem that traditional knowledge integration needs to be processed through an additional network or branch is solved, the needed calculation amount and memory overhead are saved, BAKE is applied to a network architecture, the classification performance of the network architecture can be improved, and wrong supervision caused in the processing process is avoided.
Owner:FOSHAN NANHAI GUANGDONG TECH UNIV CNC EQUIP COOP INNOVATION INST +1
Toy schoolbag
The invention relates to a toy schoolbag comprising a hollow schoolbag body. A plurality of shell fabrics are connected to the surface of the schoolbag body and are internally provided with branch bags, and the branch bags are connected to the schoolbag body. In the toy schoolbag disclosed by the invention, the branch bags are arranged in the shell fabrics and the schoolbag body, and the shell fabrics are used for hiding the branch bags, thus persons are not easy to find out the branch bags; furthermore, various live patterns are drawn on the surface of the shell fabrics of the toy schoolbag, thus persons pay attention to the delicate patterns, and the branch bags are further hidden with extremely strong confusion property; in addition, the bottom of the toy schoolbag is provided with a slide fastener, an interlayer is arranged in the schoolbag body, the persons are not easy to find out the slide fastener due to traditional knowledge, and other persons can not find out a lower cavity further due to the arrangement manner of the interlayer, thus the confusion property is extremely strong.
Owner:朱俊芳
Who we serve
- R&D Engineer
- R&D Manager
- IP Professional
Why Eureka
- Industry Leading Data Capabilities
- Powerful AI technology
- Patent DNA Extraction
Social media
Eureka Blog
Learn More Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.
© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com