37 results about "Shared learning" patented technology

In various embodiments, the present invention is directed to a decentralized and secure method for developing machine learning models using homomorphic encryption and blockchain smart contracts technology to realize a secure, decentralized system and privacy-preserving computing system incentivizes the sharing of private data or at least the sharing of resultant machine learning models from the analysis of private data. In various embodiments, the method uses a homomorphic encryption (HE)-based encryption interface designed to ensure the security and the privacy-preservation of the shared learning models, while minimizing the computation overhead for performing calculation on the encrypted domain and, at the same time, ensuring the accuracy of the quantitative verifications obtained by the verification contributors in the cipherspace.

The invention relates to a method for quickly classifying and retrieving sentences in an article by using an electronic device. The electronic device generates a classified catalog listing of the sentences in the article according to a specified classification method; when in retrieval: a processor extracts the sentences in the content of an e-book opened by the user one by one, searches the classified catalog in which the sentence is positioned, and adds the classified catalog name to the sentence as a classification mark annotation; and after the sentence with the classification mark is selected by the user, a sentence reading pointer is positioned in the classified catalog, and the sentences therein are output. The method provides a brand-new learning mode for people to read excellent classic articles. The user can annotate the contents in which the user is interested in the learning process, and build an own reading notes database of the user. The reader also can modify and improve a ready-made material database according to own preference, learns by sharing with others through the Internet, and improves the reading and learning efficiency in cooperative learning.

The invention relates to the technical field of machine learning, and aims to provide a neural network model training method based on shared learning. According to the method, a plurality of shared learning participants perform hybrid security multi-party calculation and central node local calculation through secret sharing to realize neural network model training; the method specifically comprises the steps of performing data preparation; loading and initializing data; loading the current training batch; calculating a first-layer linear output; calculating a model prediction value; calculating a model output gradient; updating the model and calculating; updating the local sharing weight; performing assessment on a test set. According to the method, multi-party security calculation and local calculation are combined, and the shared learning efficiency is improved. Through an online encryption multiplication triad provider, the efficiency of secure multi-party calculation of secret sharing is improved, and the overhead is greatly reduced compared with schemes of homomorphic encryption, confusion circuits and the like. The method is suitable for a longitudinal shared learning scene,and can enable all parties to carry out federated modeling under the condition that no overlapping features exist.

The invention mainly belongs to the technical field of multi-target optimization, and particularly relates to a novel multi-target particle swarm optimization method. The method is based on shared learning and Cauthy mutation. Shared learning factors are adopted in the method to change the speed and position updating formula of particles, and the optimal position of the particle individuals and external files are updated by adopting Cauthy mutation operators. According to the method, the global search capability and the local optimization search precision of the particles can be enhanced, and the algorithm is enabled to fast approach Pareto leading edge and premature convergence of the algorithm can be avoided. According to the method, solution convergence, diversity and distributivity of the multi-target particle swarm algorithm for processing the problem of multi-target optimization can be enhanced.

The invention provides a robot shared learning system based on a cloud platform. The system comprises a private model generation terminal and a cloud fusion computing sharing platform. The private model generation terminal is used for uploading the locally generated private model and the environmental feature information collected for generating the private model to the cloud fusion computing sharing platform. The cloud fusion computing sharing platform comprises a model fusion computing module. The model fusion calculation module is used for combinding the uploaded private model with other robot terminals, a private model generation terminal and / or other private models to generate uploaded environment feature information and merging and calculating the information with a sharing model stored on a cloud sharing computation sharing platform to generate a novel sharing model. The novel sharing model is downloaded and / or studied by other robot terminals, the private model generation modeland / or other private model generation models.

Under the present invention a hierarchical tree corresponding to the learning environment is provided. When the state of a shared learning competency in the learning environment is changed, branched rollup through the hierarchical tree is performed so that all applicable nodes are rolled up without a node being rolled up more than once. Specifically, when the state of the shared learning competency is changed, control blocks are generated for each predecessor of the “sharing” nodes. The control block for each predecessor identifies the successors (nodes) that must be rolled up before the predecessor itself can be rolled up. Under this methodology, a node will only be rolled up when all necessary successors have been rolled up, thus, preventing repeated rollup of any of the nodes.

The present invention provides a method, system and computer program product for administering learning objects within a learning management system. In the invention, a learner can select various learning materials in the system, known as “learning objects”, and can aggregate these learning objects in a special category or “learning folder” directly on the system. The learning folders can be separate from an existing course catalog of learning objects. Each learner can establish and delete learning folders, and can modify each learning folder by adding or removing learning objects to and from the folder. Finally, while each learning folder created by a user can at first only be accessed by the user, a user can allow other learners in the system to access the folder by establishing a “learning link” for the folder, which dictates which other learners can access the learning folder.

The invention discloses a creation sharing and learning management system and method. The system comprises a data input generation module used for inputting staff information and automatically generating staff login passwords and credit points needed to be finished by staffs, a login module used for login, a video uploading permission management module used for submitting video uploading permission application data, performing auditing, and if the auditing is passed, obtaining a video uploading permission, or otherwise, not obtaining the video uploading permission, a video recording and uploading module used for recording and uploading a learning video, a storage module used for storing the uploaded learning video, and a video management module used for auditing the uploaded learning video, and if the auditing is passed, releasing the learning video, or otherwise, not releasing the learning video. Knowledge of front-line leaders and management elites in a company can be recorded through a mobile video technology and is shared and learnt with the staffs in the company through a platform, so that very precious knowledge wealth is accumulated for an enterprise and the training education cost is reduced.

The invention relates to the technical field of network equipment application, in particular to a network education shared learning system which comprises a base, an insertion rod, a protective shell,a tablet personal computer body and a fixing mechanism. The insertion rod is arranged at the top of the base. The protective shell is fixedly connected to the top of the insertion rod, and the tabletpersonal computer body is arranged in the protective shell. A cavity is formed in the insertion rod. The fixing mechanism is arranged in the cavity. The fixing mechanism comprises an electric push rod, a first push plate, a first air bag, an air guide pipe, a shell, a second push plate, a push rod and a pressing plate. The electric push rod is arranged at the top of the cavity. The first push plate is arranged at the output end of the electric push rod. The shell is fixedly connected to the bottom of the cavity. The first air bag is arranged at the top of the shell. The air guide pipe is arranged on the side wall of the first air bag, and the first air bag communicates with the shell through the air guide pipe. A tablet personal computer can be stably mounted at the top of the base.

The invention provides a model processing method for shared learning. The method comprises the following steps: obtaining a plurality of candidate model update values for an initial neural network model; performing quality evaluation on the plurality of candidate model update values to obtain weights of the plurality of candidate model update values; adopting a median mechanism with weight for the plurality of candidate model update values, and determining a model update value for updating the initial neural network model; and updating the initial neural network model based on the determined update value. By firstly obtaining the plurality of candidate model update values, then obtaining weights of the plurality of candidate model update values through quality evaluation, and then determining the model update value for updating the initial neural network model based on the weighted median mechanism, based on the weighted median mechanism, it can be avoided that a false trained neural network model is adopted to update the initial neural network model, and the problem that an updated neural network model does not conform to a real situation by adopting an existing model updating method can be solved.

Disclosed are a smart learning system using a device social relation and a method thereof. The smart learning system includes a first device that generates a device social relation and shares learning content, and at least one second device that executes an event related to the learning content by acquiring the shared learning content when is included in the device social relation.

The invention discloses a campus mutual assistance sharing platform based on a WeChat applet and an operation method thereof. The platform comprises a front-end management system and a background database, the front-end management system comprises a home page unit, a message unit and a personal data unit. The home page unit comprises a home page display page and a learning sharing sub-page arranged on the home page display page; the learning sharing sub-page comprises a question and answer module, a resource sharing module, a class rubbing module and a science and technology frontier module; wherein the question and answer module is used for asking questions and answering questions among users, the resource sharing module is used for sharing learning and living resources among the users, the class rubbing module is used for the users to query school course information, the science and technology frontier module is used for the users to browse school scientific research information, andthe background database is used for storing the school course information and personal information of the users. According to the invention, full utilization of platform resources can be promoted, relevance and interactivity among students in schools are improved, the learning atmosphere of students in schools is aroused, and the communication circle of students is expanded.

The invention discloses a low-illumination pedestrian detection method and system based on multi-task feature fusion shared learning. The method comprises the steps: acquiring normal and low-illumination pedestrian data sets; pre-training an image illumination enhancement network by using the normal and low-illumination pedestrian data sets; pre-training a pedestrian detection network by using thenormal illumination pedestrian data set; designing a multi-task feature fusion module capable of fusing features between upstream and downstream tasks, performing feature fusion and sharing on the two networks, and constructing a low-illumination pedestrian detection network based on multi-task feature fusion shared learning; importing the two pre-training models into the low-illumination pedestrian detection network, and performing training by utilizing the normal and low-illumination data sets to obtain a low-illumination pedestrian detection model based on multi-task feature fusion sharedlearning; and detecting a detected image by using the low-illumination pedestrian detection model based on multi-task feature fusion shared learning to obtain the position of a pedestrian in the image. According to the method, the position of the pedestrian can be accurately and efficiently detected in the low-illumination image.

The chemical mechanical grinding time setting method based on clustering and multi-task learning belongs to the field of automatic control, information technology and advanced manufacturing. The method takes the process index and product state as input, and the key factor affecting the process index - grinding time is used as the output to establish a reverse model to optimize the setting of grinding time. In the process of constructing the above reverse model, in view of the problem of many types of production varieties and little data of a single variety, similar varieties are clustered according to product characteristics, and a multi-task learning method based on common parameter extraction is used for modeling in each category; The calculated model parameters are divided into the common part of this kind of variety and the private part of single variety.

In various embodiments, the present invention is directed to a decentralized and secure method for developing machine learning models using homomorphic encryption and blockchain smart contracts technology to realize a secure, decentralized system and privacy-preserving computing system incentivizes the sharing of private data or at least the sharing of resultant machine learning models from the analysis of private data. In various embodiments, the method uses a homomorphic encryption (HE)-based encryption interface designed to ensure the security and the privacy-preservation of the shared learning models, while minimizing the computation overhead for performing calculation on the encrypted domain and, at the same time, ensuring the accuracy of the quantitative verifications obtained by the verification contributors in the cipherspace.

The invention discloses an evacuation simulation method and system based on multi-agent shared Q learning, wherein an evacuation scene is bulit, a crowd evacuation doble-layer control mechanism is built, the upper space is composed of management agent, navigation agent and knowledge base, and the lower space is pedestrian to be evacuated; the crowd is grouped, each group selects the leader, establishes the association between the group and the navigation agent, and the leader of each group corresponds to a navigation agent; each navigation agent puts real-time information into the bulletin board managed by management agent in the process of evacuating each group; multi-Agent adopts shared Q learning algorithm to learn, and the result of shared learning is sent to navigation Agent in real time, so that navigation Agent can plan the route according to the global evacuation situation. Each group evacuates to the evacuation exit according to the guidance of the corresponding navigation agent.