42 results about "Deep knowledge" patented technology

The invention discloses a neural network and fuzzy control fused electrical fire intelligent alarm method. The method comprises the following steps of: 1, acquiring a leakage current signal, current and voltage signals, an arc light signal, a temperature signal and a field electromagnetic environment parameter signal by using a sensor on site, and pre-processing signals acquired by the sensor by using a velocity detection algorithm; 2, transmitting processed data to a three-layer feedforward error counterpropagation neural network and processing, wherein the neural network is subjected to supervised learning and establishes a weight matrix in advance; and 3, transmitting electrical circuit undamage probability, electrical circuit damage probability, and electrical circuit fire probability output by the neural network to a fuzzy inference module and performing fuzzy inference to acquire a forecast result of electrical fire. In the method, the probability of the electrical fire is accurately forecast by using the advantages of advanced theories, such as neural network, fuzzy control and the like, and without depending on deep knowledge of an object, the electrical fire forecasting accuracy is obviously improved and the damage of the electrical fire can be effectively prevented and reduced.

The invention discloses a system and method for matchmaking and transaction of electronic commerce. The system comprises a user management module, a target management module, a transaction management module and an intermediary module, wherein an intermediary user provides commodities, services or user recommendation service for a buyer or a seller through the intermediary module; the commodities and services needing to be purchased or user parameter information provided by the buyer and the commodities, services or the user parameter information provided by the seller are received and stored into a database by the intermediary module; the commodities, services or user parameter information needed by the buyer and the commodities, services or user parameter information needed by the seller are matched, and the matching coincidence rate is calculated; and a group of commodities, services or user data or multiple groups of commodities, services or user data are selected for the two parties of the buyer and the seller to select according to the matching coincidence rate. According to the system and the method, the intermediary user who has deep knowledge of certain category products can effectively help the two trading parties reduce information query time and cost; and the success rate of on-line transaction is greatly improved due to the appearance of the intermediary user.

The invention discloses a domain knowledge modeling and knowledge level estimation method in an adaptive learning system, and particularly relates to the field of adaptive learning. The problem that class education services in the existing education field are difficult to meet personalized education requirements is solved. The method specifically comprises two aspects of domain knowledge modelingbased on a knowledge graph and knowledge level estimation based on a deep learning technology. The domain knowledge modeling based on the knowledge graph comprises the steps of determining a subject knowledge graph framework; Collecting digital teaching resources, analyzing structural characteristics and language characteristics of texts of the digital teaching resources, and constructing an education corpus used for extracting subject knowledge graphs, wherein the education corpus comprises four corpora of word segmentation and part-of-speech tagging, syntactic analysis, named entities and entity relations; Subject knowledge graph construction. Knowledge level estimation based on a deep learning technology comprises the steps of domain data preprocessing and vectorization, automatic extraction of domain features by adopting a CNN, feature vector dimension reduction, and deep knowledge tracking based on an LSTM model.

The invention relates to the technical field of deep learning, and provides a question recommendation method and device based on deep learning and machine learning, and a storage medium, and the method comprises the steps: obtaining the historical question-making data of a user, and inputting the historical question-making data into a deep knowledge tracking model, so as to obtain the mastering degree data of the user for each knowledge point; sorting the knowledge points according to the mastery degree data of the user for the knowledge points and the weights of the knowledge points, and obtaining target knowledge points according to a sorting result; based on the mastery degree data of the user for the target knowledge point, obtaining the correct answering probability of the user for each question under the target knowledge point through a parameter logarithm model; and recommending questions to the user based on the correct answering probability. According to the invention, the true mastery degree of students to knowledge points is calculated through the historical records of the students doing questions, and personalized question recommendation is provided for the students.

The embodiment of the invention discloses a face feature point detection method, device and equipment based on deep knowledge migration, and the method comprises the steps: providing a human face dataset, and carrying out the cutting of a face image according to a face detection frame provided by the face data set or a surrounding frame of face feature points, and obtaining a training set, a verification set and a test set; inputting the test sample and the training sample into an initial face alignment network framework; training a teacher network and a student network in the initial face alignment network framework by using Pytorch, and generating a training model until the loss function and the maximum number of iterations meet a predetermined condition; freezing model parameters of the teacher network, extracting deep-layer dark knowledge learned by the teacher network, and transmitting the deep-layer dark knowledge to the student network to generate a final face alignment networkmodel; and inputting the RGB face image in the natural scene into the final face alignment network model, and outputting a face feature point detection result. The face feature point detection precision is high, and the model parameter quantity and the calculation complexity are low.

The present invention relates to a method for transmitting transactional commands and data between computer networks, and in particular a method for securely transmitting transactional commands and data between two computer networks by encoding the transaction based upon the ontology and semantics of the transaction commands. The invention establishes the process in which transactional commands are decomposed into context less text, based upon a local representation of the transaction command ontology and semantics. The invention also establishes the process for authenticating the transaction using deep knowledge of the user and business rules for the transaction. The decomposition of the transaction command and data is achieved using command semantic knowledge objects, data semantic knowledge objects and a decision cube populated as a result of an examination of the transaction command set ontology and semantics. The resultant transmission consists of a set of indexes or tags and context less data fields.

The invention relates to data analysis, and provides a question recommendation method, which comprises the steps of collecting historical answer data of a student side, and predicting a first masterydegree of the student side for learned knowledge points through a deep knowledge tracking model; according to the learning time of the learned knowledge points and the forgetting curve, obtaining theforgetting degree of the knowledge points by the student side; combining the first mastery degree and the forgetting degree of the knowledge points by the student terminal to predict the second mastery degree of the learned knowledge points by the student terminal; extracting question related information according to the historical answer data of the student terminal and the second mastery degreeof the learned knowledge points; predicting a first correct rate of the student terminal on the condition that the student terminal does not do the question according to the Bayesian model through thequestion related information; sorting the questions which are not done according to the first correct rate, and selecting a first set number of questions sorted in the front to push to the student side. The invention further provides an electronic device and a storage medium. Different questions are recommended according to the knowledge point mastery degree of different students.

The invention relates to a method for querying a hierarchical semantic venation document, which is applied to the query of an unstructured document and can carry out complicated query and deeper knowledge acquisition, such as network document query, questionnaire survey document query, and the like. In the method, the query intention of a user is expressed by a query tree, the query semantic venation is expressed by a leaf node of the query tree, namely a query clause, and a logical operation relation among subnodes is expressed by a branch node. When whether a document is the document needed by the user or not is tested, the deep search and test are carried out according to the query tree. If a node is a query clause node, a semantic venation matching test is carried out; if the node is the branch node, the logical operation is carried out. The current node decides a self return value according to a return value of the subnode. Finally, if the return value of a root node is true, the current tested document is selected as the document needed by the user.

The invention provides a crowdsourcing worker performance prediction method based on deep knowledge tracking. The crowdsourcing worker performance prediction method comprises the following steps: 1, data preprocessing: converting a task data form of a knowledge-intensive crowdsourcing task into data which can be used for training a knowledge tracking model, and specifically dividing the data intotask result binarization and data splitting; 2, model training: training a deep knowledge tracking model by using the converted data to obtain a worker performance prediction model; and 3, predictingthe performance of the workers, namely predicting the performance of the workers by using the trained model.

The invention provides a practical effect evaluation and learning path recommendation method based on cognitive diagnosis. The practical effect evaluation and learning path recommendation method comprises an expansion feature preprocessing model, a deep knowledge tracking model, a knowledge network construction model and a path recommendation algorithm based on cognitive ability. The expansion feature preprocessing model performs primary evaluation of cognitive competence according to skill attributes in a learner test process, and introduces personalized difference information into the diagnosis model. And the deep knowledge tracking model predicts the knowledge mastery ability of the learner according to the test sequence and the heterogeneous features of the implicit coding, and the knowledge mastery ability serves as a basis of learning guidance. The exercises and knowledge network construction model provides a global map of scientific thinking, and the cognitive diagnosis is combined to recommend a learning path to a learner, so that the cognitive ability difference in the learning process is considered, and the logic of a knowledge structure is followed.

The invention discloses a knowledge tracking method and system based on federated learning, and the method comprises the steps: employing a deep knowledge tracking model, carrying out the modeling ofthe knowledge state of a local student through combining with a recurrent neural network, and calculating the confidence of local data; finally, enabling the server take the confidence coefficient asa weight; and weighting and averaging the parameters of each model to obtain a global knowledge tracking model. The private data of each client is better utilized, the data privacy is kept, the ability of modeling students is finally tracked more accurately, and the defects of privacy information leakage, neglect of education data quality and the like of the existing method are overcome.

The invention discloses a student Bloom mastery degree evaluation method and system and a storage medium. The method comprises the steps: constructing and training a deep knowledge tracking model; when the student generates a new learning behavior, predicting the question solving accuracy of the student on each exercise through the trained deep knowledge tracking model; and evaluating the Bloom grasp degree of the knowledge point of the student at the current moment based on the prediction result of the question solving accuracy of the student on each exercise. The Bloom mastery degree of theknowledge points can be modeled through the learning behavior sequence of the students and the deep knowledge tracking method, the ability information and the mastery degree of the students on the knowledge point granularity can be visually represented, and the method and system can be deployed to an online education and mixed education platform. Study ability changes and cognitive state changes of students are effectively represented, and the purpose of Bloom operation is achieved.

The disclosure describes methods and apparatus for quickly prototyping of a solution developed using one or more sensing devices (e.g., sensors), functional blocks, algorithm libraries, and customized logic. The methods produce firmware executable by a processor (e.g., a microcontroller) on an embedded device such as a development board, expansion board, or the like. By performing these methods on the apparatus described, a user is able to create a function prototype without having deep knowledge of the particular sensing device or any particular programing language. Prototypes developed as described herein enable the user to rapidly test ideas and develop sensing device proofs-of-concept. The solutions produced by the methods and apparatus improve the functioning of the sensor being prototyped and the operation of the embedded device where the sensor is integrated.

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.

The invention discloses a learning path optimization method based on deep knowledge tracking and reinforcement learning, and belongs to the field of adaptive learning. The method specifically comprises the following steps: for a certain student, selecting all unlearned knowledge points and knowledge points which are not repaired first as knowledge points to be selected; and carrying out one-hot coding by utilizing knowledge points of historical learning, inputting the coded knowledge points into the DKT model, and outputting a mastering level prediction value of each knowledge point to be selected. Then, selecting the knowledge point K with the highest prediction result and recommending the knowledge point K to students for learning; achieving the learning process by using a learning path optimization algorithm in the knowledge points; after the current knowledge point K passes learning, judging whether subsequent knowledge points exist or not, and if yes, adding the subsequent knowledge points into the knowledge point set to be selected, and moving out the current knowledge point K; otherwise, directly moving out the current knowledge point K, and selecting the next knowledge point to predict and learn again until the knowledge point set to be selected is empty. According to the method, the recommendation precision can be greatly improved, and the efficiency is improved under the condition of obtaining the same learning effect.

The embodiment of the invention discloses a picture dynamic search method and device, equipment, a server and a storage medium. The method includes: obtaining a picture of a target object by searching; determining an action of the target object, and sending an image sequence acquisition instruction to the server according to the action. The image sequence acquisition instruction is used to instruct the server to acquire at least two frame sequence pictures required for the target object to perform the action; receiving and displaying the at least two frame sequence pictures returned by the server, so that the target object performs the action in a search result page. The method dynamically displays the search result to the user in the form of multiple consecutive image frames on the user end through interaction with the user and improves the transmission speed and display efficiency of search results, so that the 3D model of the server is copyright protected; at the same time, the usercan perceive interactive feedback to enhance the user's deep knowledge of the search results.

Digital learning or tutoring systems as described herein embed, by a trained machine learning knowledge tracing engine, an array for learner interactions X into a static representation ej corresponding to a prior learner interaction xj and determine a contextualized interaction representation hj based on this. Digital tutoring systems described herein calculate, by a masked attention layer of the trained machine learning knowledge tracing engine, an attention weight Aij based on a time gap between two learner interactions with the system, and can calculate a contextualized interaction representation hj, wherein the contextualized interaction representation hj is proportional to the attention weight Aij. The systems can provide for display at the GUI a second question item based on the contextualized interaction representation hj, the second question item corresponding to a recommended learner recommendation.

The invention relates to a dynamic perception test question recommendation method and system fused with deep knowledge tracking, and belongs to the field of computers. The method comprises the following steps of: 1, inputting test question interaction sequence data of students into an embedding layer, converting large-scale sparse features into low-dimensional dense features suitable for model input, and interactively sequencing test questions of the students according to time; 2, modeling a dependency relationship between test question interaction sequences by using a gating loop unit, and introducing an auxiliary loss function to supervise learning in a current hidden state by utilizing next test question interaction; 3, designing a gating circulation unit with a self-attention updating gate to model a knowledge evolution process related to the target test question; and 4, connecting the dynamic knowledge state of the student and the embedded vector of the target test question in series, and feeding into a multi-layer perceptron to carry out final Softmax judgment to determine whether the target test question needs to be recommended to the student or not. According to the method, the performance of test question recommendation prediction is effectively improved, and personalized test question recommendation for students is realized.

According to the adaptive learning support device and method based on deep knowledge tracking provided by the invention, through an improved knowledge tracking model, the current knowledge state of a learner is tracked in real time in a relatively accurate and practical manner. The device comprises an improved deep knowledge tracking model, a mastered knowledge point topic recommendation module and a knowledge level transition resource recommendation module. According to the invention, learning behaviors of students are fused into a knowledge tracking model, and an improved knowledge tracking model is provided based on an MANN. According to the method, the knowledge mastering degree of the student can be obtained based on the knowledge tracking model, and based on the trained knowledge tracking model, unfinished questions associated with mastered knowledge points are recommended to the student, and resources based on a knowledge level transition trajectory are recommended. According to the invention, the accuracy of the knowledge tracking model is improved, and the purpose of recommending learning resources for students based on collaborative filtering is achieved.

The invention provides a mobile Internet ubiquitous education system, comprising a problem editing system module, a problem classification publishing system module, a problem distribution system module, and a client. The working principle is that the problem editing system module is used to edit problems, the problem classification publishing system module is used to publish edited problems in classification, and the problem distribution system module is used to publish problems to the client. Through the problem editing system module publishes problems in classification through the problem classification publishing system, and the problems can be checked on the client. When a student comes across a problem which cannot be solved by the student, the student submits the problem on a platform, and students or teachers would answer the problem, and the student can find related materials on the platform. The student asking the question learn knowledge, and the student answer the question consolidate the knowledge, achieving double wins. The system provides convenience for communication among students and teachers and students, and students can exchange some deep knowledge through the platform.

The invention discloses an infant early education robot, and relates to the field of intelligent early education robots. The infant early education robot comprises a head, a display screen, a microphone, operation keys, a USB data and charging port, a main body trunk, arms, touch sensors, feet, a loudspeaker and a rotary knob switch, wherein display screen and the loudspeaker are embedded into thehead; the microphone, the operation keys, the USB data and charging port and the rotary knob switch are embedded in the main body; and the touch sensors are embedded into the arms on the left and right sides. The robot has an intelligent voice interaction system, can be controlled by voice, such as playing a song, a story, an ancient poem and the like, and can also swing along with rhythm; and parents can also guide more and deeper knowledge, words and encyclopedia through the USB data of the robot according to interests and hobbies and accepting ability of a child, and the robot can teach the child in the next period of time. In addition, the robot also has the function of monitoring in real time, can accompany the growth of the child to play with the child and also can be used as an intelligent control center for the family, so that the situation of the child can be mastered by parents working outside at anytime, and a lot of technological experiences are brought to adults in leisure time is achieved.

The invention belongs to the technical field of cement clinker heat consumption, and discloses a method for constructing a knowledge graph of heat consumption in cement clinker production by a new type of dry process. Apply the relationship between knowledge to build a structured knowledge map of heat consumption in the production of new dry process cement clinker; use the distributed agent model to design agents and corresponding sub-agents based on the three-level characteristics of basic knowledge, in-depth knowledge and comprehensive application knowledge , enabling users to acquire relevant knowledge by assigning instructions to agents for execution. The present invention analyzes the heat consumption from three levels: the basic level, the depth level and the application level, discovers the potential relationship between entities, and then finds out key factors affecting heat consumption and effective and efficient methods for reducing heat consumption. So as to achieve the direct purpose of reducing energy consumption, and finally realize the optimization of cement industry.

The invention provides a deep knowledge tracking method and system, which can convert knowledge contents on a paper textbook into digital knowledge data in an editable form, and deep mining and extension are carried out on knowledge point contents by utilizing a corresponding knowledge database and a character semantic recognition result of the digital knowledge data, so that a deep knowledge tracking effect is achieved. And finally, permutation and combination are carried out on the associated knowledge point data to obtain the deep knowledge graph, so that targeted mining expansion can be carried out on the textbook knowledge content, and the deep knowledge graph based on a certain knowledge point content is formed. Therefore, corresponding divergent teaching of knowledge points in the teaching process is facilitated, and the teaching quality and the divergent thinking ability of students are improved.

A demonstrator for position relation between two circles is a teaching aid. Wherin a rectangular iron sheet flat is affixed with a sketch paper; a fixed circle is drawn on the sketch paper; another movable circle is disposed; a piece of magnet is disposed on the back side at the center of the movable circle, while fixed on a long cord; a bobbin winder and a U-shape positioner are respectively disposed at the middle of both short sides of the rectangular iron sheet flat; both ends of the long cord are tied to the bobbin winder wheels; and the long cord is configured through the center of the fixed circle. The advantages are that: by employing the demonstrator for position relation between two circles, students can intuitively know the plurality of relations between two circles in the circle motor process of the demonstrator, and find out the relationship between the distance of the two circle centers and the radiuses of the two circles in different position relationships. The demonstrator is convenient in operation and vivid, can leave a deep knowledge memory for students.

The invention provides a knowledge tracking method based on learning migration, and relates to the technical field of education data mining. According to the method, a learning curve and an Ebbinghaus forgetting curve are used as priori to track the change of knowledge proficiency and abstract principle mastering degree of students along with time, the knowledge structure of the students is divided into two parts, one part is a deep knowledge structure obtained by understanding the abstract principle by the students, and the other part is an abstraction principle. The other part is a shallow knowledge structure representing shallow association between knowledge concepts, the model performance can be effectively improved, a Q matrix marked by an educational expert is used for describing the inclusion relation between the knowledge concepts and questions, the Q matrix serves as prior to generate a question embedding result, a probability matrix decomposition framework is designed, and the framework is divided into the following three parts: a first part, a second part and a third part; student knowledge proficiency and abstract principle mastering degree tracking, student practice feedback modeling and model parameter learning, and student knowledge mastering degree tracking through a learning migration-based multi-level knowledge tracking model.

A Real-Time Group Intelligence Creation System 100 comprising of Group Intelligence Creation Controller 200, Facilitator Expert System 300, Adaptive Group Intelligence Mining Engine 400, Intelligent Web Communicator 500, Idea and Solution Source Data Server 600, and Classification, Extraction, Thinking Pattern and Hint Data Server 700. The System extends traditional computational grid to include idea creations and problem solving generations to form a collaborative thinking grid that is made up of mass volume of participants using either mobile device or stationary device, and is without the need of face-to-face interaction. The System uses both shallow knowledge and deep knowledge mining agents to mine unstructured ideas and solutions in real-time for unifying multiple topics and generating classifications, extractions, thinking patterns and hints. This information are provided to participants during creation processes in order to simulate and accelerate participants' thinking further. The System's group intelligence creation processes is without time limited and eliminates the need of using paper or board medias. The System allows participant to participate simultaneously in multiple topics, and to switch between thinker role and viewer role during creation processes.

The invention belongs to the technical field of heat consumption of cement clinker. The invention discloses a construction method of a novel dry-process cement clinker production heat consumption knowledge graph, which comprises the following steps: by taking cement clinker production heat consumption knowledge as a center, constructing a structured novel dry-process cement clinker production heat consumption knowledge graph by dividing the relationship among basic knowledge, deep knowledge and comprehensive application knowledge; adopting a distributed agent mode, designing agents and corresponding sub-agents on the basis of three-level characteristics of basic knowledge, deep knowledge and comprehensive application knowledge, and distributing instructions to the agents for execution, so that a user obtains related knowledge. According to the method, the heat consumption is analyzed from the three levels of the basic level, the depth level and the application level, the potential relationship between entities is found, and then key factors influencing the heat consumption and an effective and efficient method for reducing the heat consumption are found out, so that the direct purpose of reducing the energy consumption is achieved, and finally optimization of the cement industry is achieved.

The invention discloses a deep layer data processing method combined with a knowledge base. The method comprises the following steps that concept tuple sets in a body base are merged; link routes of predicates in different types and corresponding logic rule sets are obtained; the logic rule sets are screened preliminarily, candidate rule sets are obtained; a deep layer probabilistic graphical model is obtained; structuring tuples are obtained based on data to be processed and are mapped to a layering concept space; target tuples are generated, semantic extension is carried out; logic rule sets and evidence tuple sets are obtained; a Markov logical net is subjected to instantiation, the conditional probability for target tuple founding is computed, and data processing results are obtained. The invention further provides a deep layer data processing system which comprises a structuring module, a conceptualization module, a target generating module, an extension module, an activating module and a probability computing module. Context and background knowledge can be fully merged, and accordingly, the purpose of semantic comprehension is really achieved.

The invention provides a method and device for determining the mastering degree of knowledge points and electronic device.The method comprises the steps that based on questions in multiple pieces of test paper and answer data of a user for the questions in the multiple pieces of test paper, the mastering hierarchy of the user for knowledge points corresponding to the questions in each piece of test paper is determined; and inputting the mastering hierarchy of the knowledge points corresponding to the questions in each test paper into the deep knowledge tracking model, and determining the current mastering hierarchy of the user for the knowledge points based on an output result of the deep knowledge tracking model. According to the method and device for determining the mastering degree of the knowledge points and the electronic equipment provided by the embodiment of the invention, questions with a plurality of knowledge points can be processed, test papers with comprehensive questions with a plurality of knowledge points can be selected according to teaching requirements, the workload of teachers is reduced, and the teaching efficiency is improved. And the mastering degree, namely the current mastering degree, of the students on the knowledge points in a certain stage can be accurately and efficiently determined.

The invention discloses a chicken MDA5 gene promoter, the sequence of which is shown as SEQ ID NO: 3 in the sequence table, and the length of the MDA5 gene promoter sequence is 2487bp. The invention also discloses the cloning method and application of the MDA5 gene promoter. The invention verifies the ability of the chicken MDA5 gene promoter to initiate gene expression in cells, verifies the regulation and control of the promoter expression by different stimuli, and brings deeper cognition for the expression regulation of the MDA5 gene.