40 results about "Collaborative knowledge" patented technology

The present invention discloses a computer-implemented online-offline workspace and method for creating, developing, storing, and managing digital content within a contextual and shared knowledge network. The invention includes a central service facility that provides an online platform for the users to work in a context-based and shared knowledge environment through a user interface on a wide range of user access devices. The online platform is embedded with a plurality of applications to allow the user to capture, create, develop, store, process, share, distribute, retrieve, reuse, and manage digital contents containing any one or a combination of the following: text, graphics, audio, video, whole or portions of web-pages and web-links. The invention further includes an end-user facility providing an offline platform that gets synchronized with the online platform upon detection of a secured communication network.

The invention enables multimodal cognitive communications, collaboration, consultation and instruction between and among heterogeneous networked teams of persons, machines, devices, neural networks, robots and algorithms during various stages of medical disease management, including detection, diagnosis, prognosis, treatment, measurement, monitoring and reporting. The invention enables both synchronous and asynchronous multiparty collaboration with multichannel, multiplexed streaming imagery data, including interactive curation, multisensory annotation and metadata tagging, as well as multi-formatted encapsulation, saving and sharing of collaborated imagery data as packetized augmented intelligence. The invention acquires both live stream and archived medical modality imagery from network-connected medical devices, cameras, signals and sensors, as well as multiomic data [phenotypic, genomic, metabolomic, pathomic, radiomic, radiopathomic and radiogenomic] maps and clinical data sets from structured reports and clinical documents, including biometric maps and movies, hapmaps, heat maps and data stream visualizations. The invention also acquires both medical and non-medical streaming imagery data from image data repositories, documents and structured reports, workstations and mobile devices, as well as from wearable computing, signals and sensors. The invention enables networked teams to interactively communicate, concurrently collaborate and bi-directionally exchange multichannel multiplexed imagery data streams, singly or together, in real time or asynchronously, generally by curating, annotating and tagging imagery information objects. The invention encapsulates and saves collaborated imagery data, together with multisensory annotations and metadata tags, in standard file formats as packetized augmented intelligence. The invention enables recursive cognitive enrichment of clinical cognitive vismemes, and saves packetized imagery information objects, multisensory annotations and metadata tags in native file formats [PDF, MPEG, JPEG, XML, XMPP, QR,TIFF, RDF, RDF / XML, SVG and DAE] as well as in formats compliant with standards for digital communications in medicine [DICOM]. The invention enables live stream multicasting of multimodal cognitive instruction and collaborative knowledge exchange with multisensory [visual, auditory, haptic] annotation of streaming imagery data, as well as secure, encrypted transmission of streaming augmented intelligence across file sharing data networks for informatics-enabled learning, specialist skills acquisition and accelerated knowledge exchange.

The invention relates to a method for carrying out personalized recommendation on commodities by fusing a knowledge graph. The invention relates to the technical field of machine learning for personalized recommendation of commodities, and the method comprises the steps: obtaining the historical behavior data of a user, and generating a commodity knowledge graph; fusing the historical behavior data of the user with the commodity knowledge graph to construct a collaborative knowledge graph; adopting a representation learning method in machine learning to obtain vector embedding representation of nodes and relationships in the collaborative knowledge graph; embedding vectors of nodes and relations in the collaborative knowledge graph and inputting historical behavior data of the user into agraph convolutional neural network model based on an attention mechanism to form vector embedding representation of new nodes and relations containing neighborhood information; and defining a score function, calculating the probability that the user likes the commodity through the user and commodity vector embedding representation of the fused neighborhood information output by the graph convolutional neural network model, and performing sorting according to a prediction result to obtain a recommendation list.

A community server categorizes and accumulates messages exchanged by users on a virtual community for respective topics. The community server includes a subscription type setting unit, which provides, to each user and predetermined administrator, a mechanism for selecting a subscription type indicating the way each user subscribes to messages posted to the virtual community. For this purpose, the subscription type setting unit manages the subscription types of respective users for each community using a subscription type table of community management information, and changes the subscription type in the subscription table in accordance with a request from each user and administrator.

A collaborative, semantic-based knowledge structure creation and management system and method are provided. The system and method may include knowledge structures which include concept based maps which may be linked according to topic, relationship types, conceptual equivalence, and temporally. The knowledge structures may be managed by users, wherein users may perform at least one of adding a new concept, creating a new relationship, confirming an existing relationship, and breaking an existing relationship.

The invention discloses a tourism route recommendation method and system based on a tourism knowledge graph. The method comprises the steps of obtaining scenic spot data and user comment data, carrying out data cleaning, and unifying data specifications; constructing an interaction information graph between the user and the scenic spot by using historical behavior data generated by the user for the scenic spot; constructing a scenic spot knowledge graph, forming a relational graph between scenic spots and scenic spot attributes, then importing data into a graph database, and displaying the constructed scenic spot knowledge graph in a visual form; constructing a collaborative knowledge graph with a high-order relationship by adopting the interaction information graph and the tourist attraction knowledge graph, and recommending tourist attractions conforming to the interest of the user through the collaborative knowledge graph; and carrying out tourist route planning on the recommended tourist attractions to obtain a complete tourist route. According to the invention, personalized scenic spot recommendation and route planning service can be provided for the user, and the recommendation result is more accurate and has interpretability.

The invention provides a personalized recommendation method based on a knowledge graph, and the method comprises the steps: S1, combining historical interaction data of users in a user set and articles in an article set and an original knowledge graph into a collaborative knowledge graph, in the collaborative knowledge graph, connecting entities corresponding to the user and the interactive article thereof in the knowledge graph by using directed edges; S2, expressing each entity and relationship in the collaborative knowledge graph by using quaternion vectors; S3, embedding the quaternion ofthe entity into a path along the collaborative knowledge graph by adopting an attention mechanism to perform preference propagation and aggregation; S4, constructing a preference score prediction function to calculate preference scores of each user and different articles on the basis of the collaborative knowledge graph after preference propagation and aggregation are completed; S5, adopting a loss function to jointly optimize an embedding and preference score prediction function of the collaborative knowledge graph; and S6, predicting the preference score of the user for the new article by adopting the optimized preference score prediction function to obtain a new article recommendation list for the user.

The invention provides a recommendation method based on a heterogeneous propagation collaborative knowledge awareness network. The method comprises the steps: carrying out the explicit coding on collaborative signals between a user and an object through collaborative propagation, and proposing a method for associating and combining the collaborative signals and knowledge. Specifically, a heterogeneous propagation strategy is adopted to explicitly encode two kinds of information, a knowledge perception attention mechanism is applied to distinguish contributions of different knowledge-based neighbors, and a new method for combining collaborative information and knowledge information is provided.

System and method related to a distributed collaborative knowledge generation system are disclosed. The distributed collaborative knowledge generation system includes one or more databases configured to store content information related to the distributed collaborative knowledge generation system, a search services module configured to search for the content information, a web services module configured to retrieve and gather the content information, a presentation services module configured to share the content data with multiple users, and a data services module configured to manage the content information by providing interfaces between the one or more databases, the search services module, the web services module, and the presentation services module.

The invention discloses a large-scale collaborative knowledge processing method and a system, wherein the method comprises the following steps: receiving problem information and background information; receiving answer information and evidence information given by participants; receiving the addition and / or modification of the answer information, the evidence information, and / or the background information so as to form different versions of answer information, evidence information and background information; according to the score values of the different versions of answer information, evidence information and background information and the reliability values of scorers of the score values, calculating the reliability values of the different versions of answer information, evidence information and background information, and selecting optimal answer information; and updating the reliability values of participants providing the answer information. According to the method and system disclosed by the invention, the needs of a 'centering-on-problems' data modeling process can be satisfied better, the development mode of the modern information technology can be adapted better, and the rules of modern knowledge base construction can be conformed better.

The invention provides a recommendation system based on a non-sampling collaborative knowledge graph network. The recommendation system comprises an embedding module which is set to obtain an initial embedding vector of a triple in a knowledge graph; a non-sampling knowledge graph convolution module which is set as a single-layer convolution network comprising a plurality of linear aggregators, and is used for performing non-sampling pre-calculation on an initial embedded vector to obtain deep information of a triple, and combining the embedded vector and the deep information as an updated embedded vector; a cooperative propagation module which is set to simultaneously encode cooperative signals in user and project interaction as initial preferences of a user and a project, and is combined with the updated embedded vector to serve as an input vector of the prediction module; a prediction module which is configured to obtain a recommendation result according to the input vector. According to the method, only by designing a relatively complex propagation matrix and pre-calculation operation, the performance which is not worse than that of a depth model, the speed is higher, and a more accurate prediction result is obtained.

A system comprises servers in network connection, wherein the servers are coupled to a data knowledge base and comprise processors executing software from non-transient mediums. The software provides interaction websites which can be accessed by users executing browser applications, the websites allowing users to upload and store shorter media lessons having respective descriptive names; categories organizing the media lessons as subjects; a mechanism allowing the users to add media files associated with a single media lesson; and a mechanism allowing the user to find the media lessons and the associated media files through the categories of the subjects and interact with the media lessons and the associated media files.

The invention relates to the technical field of data processing, in particular to a security cross-domain recommendation method based on a multi-feature collaborative knowledge graph and a block chain. The method comprises the steps that a client stores user data which is not uploaded to a server, and uploads the user data after preprocessing the user data; a multi-feature collaborative knowledge graph is constructed to perform cross-domain information security fusion; all participants register and join the block chain, and write an operation upper chain so as to carry out data anomaly detection analysis and source tracing; and the user characteristics and the knowledge graph are taken as input of a GCN, and a personalized recommendation list is output to a target user. Features are used as entity nodes in the knowledge graph to hide sensitive information associated with individuals, and privacy security is protected while user data analysis requirements in recommendation are met; for model poisoning attacks, a block chain technology is introduced to realize traceability of upper chain data, and the security of the data is ensured from the source; and the correctness and integrity of the collected data are realized, so that the recommendation accuracy is improved.

The invention relates to a knowledge graph recommendation method and system based on an improved KGAT model. The method comprises the following steps: constructing a domain knowledge graph; taking the user as a node of a domain knowledge graph, taking historical interaction between the user and an article as an edge of a relationship, adding the historical interaction into the knowledge graph, and constructing a collaborative knowledge graph; using an improved KGAT model to learn the collaborative knowledge graph to obtain vectorized representations of users and articles and weight information of adjacent edges of article nodes; sorting all the to-be-recommended article sets, and selecting the first N articles as a article set to be recommended for the user; generating a recommendation reason based on the sorting result and the weight information of the adjacent edges of the article nodes corresponding to the sorting result in the collaborative knowledge graph; and when the user requests recommendation, generating a recommendation list from the recommendation reasons and the articles, and returning the recommendation list to the user. According to the method and the system, the personalized recommendation list is generated for the user, and meanwhile, the personalized recommendation reason can be generated, so that the credibility of the recommendation result is improved.

The invention belongs to the field of man-machine interaction topic transfer, and particularly relates to a man-machine interaction topic transfer method based on a collaborative knowledge graph. The method comprises the steps that a candidate topic set and a similar user set are acquired according to the social relation and behavior history of topic participants; constructing a user-topic bipartite graph according to the candidate topic set and the similar user set; obtaining a knowledge graph, and fusing the user-topic bipartite graph with the knowledge graph to obtain a collaborative knowledge graph; inputting the collaborative knowledge graph into a man-machine interaction topic transfer model based on the collaborative knowledge graph to obtain a transfer topic; the topic participant takes the transferred topic as reply content to carry out topic switching; according to the method, two factors of similar users and similar topics are considered, the collaborative knowledge graph is constructed, the high-order relation between dialogue related information is fully mined, dialogue content features are greatly enhanced, dialogue topics are naturally and coherently switched in the man-machine interaction process, and the dialogue interest and satisfaction degree of topic participants are improved.

Provided is a system that offers user with various automated guidance which are identified intelligently based on current analysis path of the user to provide automatic recommendations for analysis path so as to ease the big data analysis. The recommendations are based on other expert user's analysis already performed. The user chooses analysis path recommendations to arrive at the final result inless time and with ease. The system can learn continuously from the analysis path of the other users on similar data. The system utilizes the collaborative knowledge form all users for giving recommendations. The system and / or an apparatus (800) has a receiving module (808), a user interaction profiler module (810), a user profile matcher module (812), and a recommendation module (814) to provideautomatic recommendations to the users.

The invention provides a medical care and health care combined knowledge service system construction method for elderly care, which supports and provides a high-precision elderly care knowledge graph by combining cross-media knowledge unified representation, cross-media knowledge entity extraction, semantic analysis and man-machine collaborative knowledge graph auditing. And an end-to-end elderly care knowledge service of risk grading, intelligent question and answer, scheme making, emergency early warning, rehabilitation and care recommendation is provided. According to the method, the cross-media knowledge graph is used for replacing a traditional relational database which is insufficient in flexibility and small in scale, and an omnibearing knowledge data set in the elderly care ecosphere is generated. Compared with an existing medical and health care service system based on a traditional database, the medical and health care service system provides more comprehensive, accurate and flexible domain knowledge integration on the basis of multi-source and multi-dimensional information, provides an information and service management scheme based on a knowledge engineering technology, provides a more automatic, personalized and intelligent elderly care solution, and improves the medical and health care efficiency. The cost can be effectively reduced; and the efficiency is improved.

The invention discloses a collaborative knowledge perception enhanced network recommendation method, and mainly relates to the technical field of computer application. Comprising the following steps: S1, constructing a knowledge graph; s2, seamlessly combining explicit coded collaborative information in interaction of the user and the project with associated knowledge in the knowledge graph, and fusing propagation information of the user and the project in the knowledge graph through a knowledge perception nerve attention mechanism; s3, constructing a cross compression unit, obtaining project original embedding through the cross compression unit, and fusing the project original embedding with knowledge graph propagation information; s4, performing score prediction; the invention can better adapt to the problems of data sparseness and user cold start, so that the model performance and generalization are improved.

A system includes an Internet connected server coupled to a data repository and having a processor executing software from a non-transitory medium, the software providing an interactive web site accessible by users executing a browser application, the web site enabling the users to upload and store relatively short core lesson files, each lesson file having a descriptive name, organization of the core lesson files into categories of subject matter, a mechanism enabling the users to add media files associated with individual ones of the core lesson files, and a mechanism enabling the users to find the core lesson files and associated media files by the categories of subject matter, and to interact with the core lesson files and associated media files.

The invention discloses a graph convolutional network recommendation method and device based on an interlayer combination mechanism, and the method comprises the steps: firstly constructing a user-project historical interaction record set, constructing a knowledge graph of a commodity, constructing a project-commodity entity alignment set, and constructing a user-project historical interaction record set; coding and integrating the interactive behaviors of the user and the project and the knowledge graph of the commodity into a unified relational graph data structure, and generating a collaborative knowledge graph of the user and the commodity; carrying out graph convolution operation on feature information of each node in the collaborative knowledge graph through a graph convolution network, and extracting user and commodity features; and finally, calculating inner products of the user and the feature vectors of the commodities in sequence, carrying out sorting recall according to the sizes of the inner products, and finally recommending the commodities to the user. According to the method, the problem of excessive smoothness in a traditional graph-based convolutional network recommendation method can be effectively relieved, and meanwhile, the method has relatively high generalization ability, so that the accuracy of a recommendation result is improved.

The invention provides a collaborative filtering recommendation method based on a knowledge graph. According to the method, a collaborative knowledge graph is obtained according to an entity interaction graph and a knowledge graph corresponding to an entity; inputting description information related to the entity into an entity description encoder to obtain an entity description vector; the entity description vector and the entity structure vector are mapped to the same semantic space through limitation of a mixed item based on a TransR algorithm, and the semantic space further comprises a relation vector; obtaining a self-center network vector of the entity according to the collaborative knowledge graph; obtaining an entity representation vector according to the entity description vector and the entity structure vector, wherein the entity representation vector is updated according to the entity representation vector and a self-center network vector; and obtaining an interaction probability between the user and the article according to the entity representation vector. According to the method, the problem of data sparseness in collaborative filtering is solved, the preference of the user and the characteristics of the article can be accurately represented, and the performance of the recommendation method is further improved.

The invention discloses a personalized recommendation method based on a knowledge graph. The method comprises the following steps: acquiring data information of a user and constructing the knowledge graph; encoding entity nodes of the mapping knowledge domain and generating embedded vector representation for each entity and relationship in the mapping knowledge domain; performing message aggregation to obtain the feature representation of the user and the feature representation of the article; and calculating the probability that each article is recommended to the user and completing personalized recommendation based on the knowledge graph. According to the method, the defect that user-project interaction records are mutually independent is overcome, and project attribute-based collaborative information of users can be obtained; the method enables the model to capture rich semantics and high-order connectivity based on inter-node feature interaction in the collaborative knowledge graph; therefore, the method is high in reliability, good in practicability and high in accuracy.

The invention discloses a memory collaborative knowledge tracking method and system. The method comprises the following steps: acquiring a question qt and an answer rt answered by a learner at a t moment; obtaining a concept matrix Mk and a concept mastering matrix at the t moment, wherein the concept mastering matrix is used for representing memory contents of the learner for each concept at the t moment in a sensory memory stage, a working memory stage and a long-term memory stage; obtaining the relevancy wt (i) of the question qt and the ith concept; calculating to obtain the total memory degree kst of the learner to the concepts related to the question qt; and outputting the prediction correct rate prt and the prediction relevancy w't of the question qt answered by the learner. The method not only comprises the representation of the memory content, but also comprises the representation of the memory stage, so that the knowledge mastering state of the learner can be more accurately represented, and the future performance of the learner can be more accurately predicted.

The invention provides a knowledge-based cooperative method of complex product development process. The method comprises five steps: (1) establishing a project task model, a research staff model and a knowledge model; (2) establishing a key word library; (3) matching cooperative tasks with research staffs on the base of knowledge; (4) perception and analysis of activity situation; (5) multi-dimensional visualization of cooperative environmental situation and related knowledge. According to the invention, knowledge is taken as a core of cooperative development process, task is taken as a drive of cooperative development process, distributions of tasks are accomplished by matching degree of cooperative tasks and research staffs, knowledge bodies and knowledge units which play supporting roles to cooperative process are acquired by analysis of cooperative situation, cooperative environment and related knowledge units are presented by multi-dimensional visualization method. The method provided by the invention has a wide application prospect in the fields of knowledge management, computer and cooperative technology.