201 results about "Dynamic learning" patented technology

A dynamic learning method and an adaptive normal behavior profile (NBP) architecture for providing fast protection of enterprise applications are disclosed. The adaptive NBP architecture includes a plurality of profile items. Each profile item includes a plurality of profile properties holding the descriptive values of the respective item. An application-level security system can identify and prevent attacks targeted at enterprise applications by matching application events against at least a single profile item in the adaptive NBP.

A system, method, and computer program product are described for dynamically assessing and improving driver performance, including that of an autonomous driving system. Based on received electronic driver performance messages, the system determines whether an instant feedback message is required to dynamically adjust driving behavior of a subject driver or autonomous driving system. If yes, then the system provides an instant feedback message, otherwise the system provides an aggregate feedback message, calibration message, and / or coaching recommendation as appropriate. An aggregate feedback message may be provided based at least on integral rating and / or driver performance score calculated and based at least on the driver performance messages received over a predetermined time period. The system may be adapted to perform one or more of the following: calculate insurance premium, insurance coverage, calculate cost of hiring a vehicle, monitor driving quality of drivers, recruit drivers, provide incentives, provide coaching, provide driver performance reports, and inform a third party about the driving performance of a subject driver.

New language constantly emerges from complex, collaborative human-human interactions like meetings—such as when a presenter handwrites a new term on a whiteboard while saying it redundantly. The system and method described includes devices for receiving various types of human communication activities (e.g., speech, writing and gestures) presented in a multimodally redundant manner, includes processors and recognizers for segmenting or parsing, and then recognizing selected sub-word units such as phonemes and syllables, and then includes alignment, refinement, and integration modules to find or at least an approximate match to the one or more terms that were presented in the multimodally redundant manner. Once the system has performed a successful integration, one or more terms may be newly enrolled into a database of the system, which permits the system to continuously learn and provide an association for proper names, abbreviations, acronyms, symbols, and other forms of communicated language.

Providing dynamic learning for software agents in a simulation. Software agents with learners are capable of learning from examples. When a non-player character queries the learner, it can provide a next action similar to the player character. The game designer provides program code, from which compile-time steps determine a set of raw features. The code might identify a function (like computing distances). At compile-time steps, determining these raw features in response to a scripting language, so the designer can specify which code should be referenced. A set of derived features, responsive to the raw features, might be relatively simple, more complex, or determined in response to a learner. The set of such raw and derived features form a context for a learner. Learners might be responsive to (more basic) learners, to results of state machines, to calculated derived features, or to raw features. The learner includes a machine learning technique.

A dynamic learning method and an adaptive normal behavior profile (NBP) architecture for providing fast protection of enterprise applications are disclosed. The adaptive NBP architecture includes a plurality of profile items. Each profile item includes a plurality of profile properties holding the descriptive values of the respective item. An application-level security system can identify and prevent attacks targeted at enterprise applications by matching application events against at least a single profile item in the adaptive NBP.

An integrated human and computer interactive data mining method receives an input database. A learning, modeling, and analysis method uses the database to create an initial knowledge model. A query of the initial knowledge model is performed using a query request. The initial knowledge model is processed to create a knowledge presentation output for visualization. It further comprises a feedback and update request step that updates the initial knowledge model.A multiple level integrated human and computer interactive data mining method facilitates overview interactive data mining and dynamic learning and knowledge representation by using the initial knowledge model and the database to create and update a presentable knowledge model. It facilitates zoom and filter interactive data mining and dynamic learning and knowledge representation by using the presentable knowledge model and the database to create and update the presentable knowledge model. It further facilitates details-on-demand interactive data mining and dynamic learning and knowledge representation by using the presentable knowledge model and the database to create and update the presentable knowledge model.The integrated human and computer interactive data mining method allows rule viewing by a parallel coordinate visualization technique that maps a multiple dimensional space onto two display dimensions with data items presented as polygonal lines.

Cached or predictive negotiation of compression parameters over a modem relay network connection involves method and apparatus for determining at a gateway the identity of a client / server modem that is called or calling, and using a profile of such called or calling client / server modem, the profile being stored at the gateway or an affiliated agent, to transmit respectively a candidate response or command containing compression parameters. Preferably, a candidate response or command includes an exchange identification (XID) or link request (LR) command or response, either provisioned, i.e. manually configured or automated, i.e. computer-generated. The profile is stored in a memory. The actual achieved parameters for a given call also may be stored, e.g dynamic learning may be had. Finally, establishing such profiles by placing dummy calls once the modem's phone number is known also is described as a feature of the invention.

The present invention relates to a method and system for efficiently determining grating profiles using dynamic learning in a library generation process. The present invention also relates to a method and system for searching and matching trial grating profiles to determine shape, profile, and spectrum data information associated with an actual grating profile.

A data compression system and method for achieving enhanced compression performance in a plurality of encoder-decoder pairs communicatively coupled in a concatenated chain is disclosed. The plurality of encoder-decoder pairs receive input data and transform the input data into compressed output data. The data compression system includes at least one quality estimator communicatively coupled to the concatenated chain to determine a plurality of quality metrics corresponding to an output from each of the plurality of encoder-decoder pairs. Further, the data compression system also includes a processing subsystem configured to optimize the operation of each of the plurality of encoder-decoder pairs based upon the plurality of quality metrics to achieve a desired level of optimization for the concatenated chain.

A method performed in a system involves, at a node within the system, receiving an input from a user, determining that the input contains an unknown word, presenting at least one response to the user, and based upon at least one additional input from the user, learning one or more associations for the unknown word.

The invention discloses an automatic sleep staging method of a single-lead electroencephalogram. A training model comprises a feature extraction module and a staging optimization module, the feature extraction module comprises CNNs (convolutional neural networks) (1) and an Softmax layer (2), the staging optimization module comprises bi-directional LSTM (long-short term memory) recurrent neural networks (3) and a CRF (corticotropin releasing factor) conditional random field model (4), and the CNNs (1), the Softmax layer (2), the LSTM recurrent neural networks (3) and the CRF conditional random field model (4) are sequentially connected. The method only needs the single-lead sleep electroencephalogram, portable and comfortable sleep monitoring requirements are met, temporal and spatial characteristics of the electroencephalogram are sufficiently excavated according to the convolutional neural networks and the recurrent neural networks, the method has dynamic learning capacity and can adapt to great changed environments of diseases, the staging optimization module sufficiently considers relation between the front and the back of N 30s of electroencephalogram data, and the staging accuracy and the generalization ability of the model are improved.

The invention discloses a cloud based smart classroom realization method and system, and relates to the field of multimedia teaching. The system comprises a cloud platform, a micro cloud server and aterminal application tool connected with one another; the cloud platform provides cloud basic facilities, a support platform, a resource service, data processing and a teaching service; the micro cloud server provides local network, storage and computing services; and the terminal application tool comprises a student terminal and a teacher terminal used for learning interaction between a teacher and a student as well as between students. The smart classroom system is realized via dynamic learning data analysis as well as cooperation among the cloud platform, the micro server and the terminal application tool, interaction between the teacher and students in class is enhanced, the class is more interesting, a learning effect of the student is analyzed and evaluated, and the student is more enthusiastic in learning and has a better learning effect.

A method for determining a facial emotion of a user in the presence of a facial artifact includes detecting Action Units (AUs) for a first set of frames with the facial artifact; analyzing the AUs with the facial artifact after the detection; registering the analyzed AUs for a neutral facial expression with the facial artifact in the first set of frames; predicting the AUs in a second set of frames; and determining the facial emotion by comparing the registered neutral facial expression with the predicted AUs in the second set of frames.

The invention relates to a network curriculum learning platform and a communication method. The network curriculum learning platform comprises a WEB server, a logic server, a database server, a content integration server, a content publishing server, a material server, a streaming media server and a network curriculum packet server, wherein the network curriculum packet server stores a network learning content packet. The network curriculum learning platform is characterized in that: the network learning content packet and the WEB server perform interactive communication through an application program interface; and the network curriculum learning platform reads basic information of a user to the network learning content packet through the application program interface on the WEB server and receives dynamic learning information of the user transmitted by the network learning content packet. By the networking learning platform and the communication method, the modularization degree is high; the network curriculum is easy and convenient to transplant; and learning resources can be shared highly.

Internet Protocol (IP) switch and method for switching an IP frame to an IP destination address. The IP switch includes access ports, interswitch ports, and a switching table. Each access port is connected to a corresponding host in a LAN. The interswitch ports are configured to be statically connected to routers within the Internet in connections that cannot be overwritten by dynamic learning. The switching table includes entries, each entry having an IP address, a port identifier, and a routing protocol internal metric indicative of confidence in the entry. The method for switching the IP frame includes: receiving the IP frame, reading the IP destination address from the IP frame, ascertaining a matching IP address in the switching table that matches the IP destination address, directing the IP frame to a specific port, transmitting the IP frame from the specific port to the IP destination address.

The invention relates to a system that assists to study language word and is implemented by a computer and a method thereof, comprising a system word test item bank, a word studying module, a module used for testing the effect of memorizing words, an analysis module of characteristics of memory, a control module of studying inferences, a word test item bank of studying inferences, an individual word bank and a dynamic study field-controlled module; the forgetting curve and the memory defects are obtained by the statistics and analysis for the personal memory process of a learner and a individual word test item bank is customized for the individual according to the two characteristics, thus improving the efficiency of assisting to study language word.

A system, method, and computer program product are described for dynamically assessing and improving driver performance, including that of an autonomous driving system. Based on received electronic driver performance messages, the system determines whether an instant feedback message is required to dynamically adjust driving behavior of a subject driver or autonomous driving system. If yes, then the system provides an instant feedback message, otherwise the system provides an aggregate feedback message, calibration message, and / or coaching recommendation as appropriate. An aggregate feedback message may be provided based at least on integral rating and / or driver performance score calculated and based at least on the driver performance messages received over a predetermined time period. The system may be adapted to perform one or more of the following: calculate insurance premium, insurance coverage, calculate cost of hiring a vehicle, monitor driving quality of drivers, recruit drivers, provide incentives, provide coaching, provide driver performance reports, and inform a third party about the driving performance of a subject driver.

Business processes are constructed and / or modified utilizing system knowledge including a folksonomy that stores information regarding a group of services usable by composition applications that collects system monitored services usage information and user inputs. In one aspect, candidate services are recommended and / or automatically updated / replaced, which can implement aspects of a business process from the group of services based upon information derived from the system knowledge. In another aspect, a composition design component interacts with a user to build and / or modify services based upon identified aspect of business processes to be implemented by services, wherein the component design component recommends candidate services that can implement the aspects of the business processes from the group of services based upon information derived from the system knowledge.

The present invention is a system and method for generating a personal voice font including, monitoring voice segments automatically from phone conversations of a user by a voice learning processor to generate a personalized voice font and delivering the personalized voice font (PVF) to the a server.

The invention provides a learning trajectory-based learning ability evaluation and extended knowledge point set recommendation method. The method comprises steps: a knowledge network and the syllabusare built; according to the syllabus, a required learning path and an initial recommended elective extended knowledge point set are built; according to each syllabus requirement and answers, a test set is built; according to the learning path and the learning state achieved by the learner himself or herself, the learner selects new knowledge points for learning, the learning condition of the knowledge points is recorded, an initial learning trajectory is built, the mastering condition of each knowledge point is evaluated, a dynamic learning trajectory for the mastering condition is built, thecomprehensive mastering ability score, the mastering breadth score and the comprehensive learning efficiency score of the learner are evaluated, and whether to achieve the current course mastering requirements is evaluated; the mastering breadth level is evaluated for the learner which already achieves the mastering requirements; and the comprehensive learning ability level of the learner is evaluated, a proper elective knowledge point set is selected, and individualized elective extended knowledge point set recommendation is carried out.

An integrated human and computer interactive data mining method receives an input database. A learning, modeling, and analysis method uses the database to create an initial knowledge model. A query of the initial knowledge model is performed using a query request. The initial knowledge model is processed to create a knowledge presentation output for visualization. It further comprises a feedback and update request step that updates the initial knowledge model. A multiple level integrated human and computer interactive data mining method facilitates overview interactive data mining and dynamic learning and knowledge representation by using the initial knowledge model and the database to create and update a presentable knowledge model. It facilitates zoom and filter interactive data mining and dynamic learning and knowledge representation by using the presentable knowledge model and the database to create and update the presentable knowledge model. It further facilitates details-on-demand interactive data mining and dynamic learning and knowledge representation by using the presentable knowledge model and the database to create and update the presentable knowledge model. The integrated human and computer interactive data mining method allows rule viewing by a parallel coordinate visualization technique that maps a multiple dimensional space onto two display dimensions with data items presented as polygonal lines.

The invention discloses a 95598 telephone traffic work order prediction and transaction early warning method based on a multi-prediction model, and relates to a power telephone traffic work order analysis method. At present, an abnormal threshold value is artificially determined through a comparison value, a loop ratio value and an amplification value, and the threshold value cannot be accuratelyand scientifically set in real time, so that the monitoring and early warning, problem positioning and trend prediction capabilities are insufficient. Based on the LSTM neural network deep learning technology, a scientific index transaction prediction model is established, the mathematical relationship of all indexes is studied, and short-term telephone traffic work order confidence transaction prediction and intelligent early warning application are achieved. According to the technical scheme, index analysis early warning is obtained from a large number of indexes more efficiently, more leanand more intelligently, and the working efficiency of customer service index analysis and quality control is improved. The defect that traditional curve fitting modeling needs periodic model correction is overcome, online real-time dynamic learning prediction and early warning analysis are supported, and the monitoring early warning, problem positioning and trend prediction capabilities of daily indexes are improved.

Business processes are constructed and / or modified utilizing system knowledge including a folksonomy that stores information regarding a group of services usable by composition applications that collects system monitored services usage information and user inputs. In one aspect, candidate services are recommended and / or automatically updated / replaced, which can implement aspects of a business process from the group of services based upon information derived from the system knowledge. In another aspect, a composition design component interacts with a user to build and / or modify services based upon identified aspect of business processes to be implemented by services, wherein the component design component recommends candidate services that can implement the aspects of the business processes from the group of services based upon information derived from the system knowledge.

The invention relates to a building method for an intelligent personalized service web site, which is a simple, easily used, highly efficient and general web site personalized service method. The method packages complex knowledge representation and knowledge acquisition processes and does not directly participate in 'personalized service', and the 'middleware' means is used for providing 'personality' of each client for an application program of the web site; the 'personality' of each client is abstracted into three aggregates which are respectively 'content preference', 'forecast preference' and 'access habit'; the data mining technology is utilized for continuous update of a knowledge base by means of establishment of a dynamic study; the 'personality' of each client is tracked and analyzed; the 'personality' of clients is provided to the application program of the web site by means of API; the application program of the web site can arrange web page contents according to the provided 'content preference' and the provided 'forecast preference' of the clients and arrange index links of web sites which can be next accessed by the clients according to the 'access habit' of the clients, thereby the aim of personalized service is achieved.

A method of updating a classifier on-the-fly is provided. The method includes providing a base classifier. The base classifier is a neural network. The method further includes receiving a class and a set of images associated with the class. The method further includes splitting the set of images into an evaluation set and a training set. The method further includes updating the base classifier on-the-fly to provide an updated classifier. Updating the base classifier includes (1) extracting features for each image from the training set from the base classifier; (2) training the updated classifier using the extracted features; and (3) scoring the evaluation set with the updated classifier.

The invention provides an intelligent immersive high-speed train virtual driving behavior evaluation method and a simulation system. According to the method, a virtual driving scene module is established, and various driving training scenes composed of multiple intelligent agents are constructed; a high-speed train dynamics module is established, including various subsystems of the high-speed train such as engine, power train system, steering mechanism and body modules; a virtual driving scene simulation module is established; a high-speed train logic simulation module is established; a data helmet display and man-machine interaction module is established, the head tracking technology of a helmet type display is researched on the basis of the virtual reality technology, and motion of the high-speed train in a virtual environment is controlled and manipulated through a high-speed operation mechanism; finally, intelligent driving behavior researching and evaluation are realized on the basis of a time varying dynamic bayesian networks (TVDBN) dynamic learning learning model according to the virtual driving operation data accumulated in a user and simulation system interaction process.

In a pressure control system having a solenoid-operated fluid valve that has an output hydraulic pressure which varies in accordance with a solenoid input signal, a dynamic learning block is configured to adjust the initial, default values for control points stored in a pressure-current (P-I) data table based on observed (measured) operating points that reflect the solenoid's actual transfer characteristic. A feed forward control block is configured to generate the solenoid input signal having a level based on the adjusted control points in the data table, which improves the accuracy of the solenoid input signal. An adjustment method uses a plurality of circular buffers each configured to store observed operating points falling within a respective range, and provides a mechanism to allow adjustment of the control points based on only partial data.

The invention provides a target arrival angle estimation method based on spatial discrete grid dynamic update, and belongs to the technical field of array signal processing. The target arrival angle estimation method is mainly for solving the problem of spatial discrete grid mismatch faced in the DOA estimation process on an array signal processing problem. Eigenvalue decomposition is performed ona signal sampling covariance matrix, through a basis pursuit method, a discrete point closest to a real angle is found, then the discrete grid points are dynamically updated by iteration, so that thediscrete grid points continuously approach a true DOA value of a target, and accurate estimation of multiple target arrival angles is carried out. The target arrival angle estimation method based onthe spatial discrete grid dynamic update effectively solves the problem of the spatial discrete grid mismatch by using a spatial discrete grid dynamic learning method, so that the algorithm can quickly estimate the DOA and keep a quite high target estimation accuracy even if the step size of a spatial discrete grid is large.