58 results about "Applying knowledge" patented technology

A GUI containing a panel along each of its edges is produced within a computer's display. Each of the four panels are always visible and accessible within the GUI and contain items that display or reference network-based information, communication, navigational systems, computer programs, and other items. These items have been assigned to the panels based upon a hierarchal containment system which, in the preferred embodiment of the invention, uses the categories “people, places, things, and home” as its first logical grouping of the various items and services a user would seek to utilize within the GUI. This organizational system, and its associated visual representation within the panels, enables the GUI to be displayed and operated within multiple computer devices, even though the designs of those devices may include a wide range of display and control systems. This consistency across various computing devices enables a user to quickly apply knowledge gained from having used the GUI on another computing device. Furthermore, the items presented within the GUI, as well as the GUI's configuration, are stored in, and retrieved from, a networked “server” computer thereby enabling a user to produce and access their familiar GUI workspace and items using a variety of networked computer devices. The GUI therefore may be accessed within multiple computing devices and provide access to information, computer programs, and electronic communications.

The invention relates to a system and method for creating, editing, using one or more multi-relational ontologies and applying knowledge contained in one or more ontologies to one or more applications. The system may extract concepts, relationships, assertions and other information from a domain specific corpus of documents. The system may also create rule-based assertions that are then curated and published as one or more ontologies that can be used in various ways.

According to a disclosed method, an MME in a network analyzes KPIs from the cells it serves and based on the KPIs, it decides to engage in sharing. The MME then contacts a sharing entity (SE) to announce that it wants to supply spectrum for sharing. The MME obtains terms of a sharing agreement from the SE and the MME obtains the identity of the other network. In response to this information, the MME configures its base stations to support the supplying of spectrum to the other network. The SE applies knowledge of network topology and of services offered. This knowledge is obtained from a sharing database. At the expiration of the sharing agreement, the SE tells the MMEs to deactivate the sharing agreement.

According to a disclosed method, an MME in a network analyzes KPIs from the cells it serves and based on the KPIs, it decides to engage in sharing. The MME then contacts a sharing entity (SE) to announce that it wants to borrow or to supply resources. The MME obtains terms of a sharing agreement from the SE and the MME obtains the identity of the other network. In response to this information, the MME configures its network elements to share. The MME's own network may be a supplier or a borrower. If the network is a supplier, the MME configures its base stations to support UEs subscribing to the other network. If the network is a borrower, the MME configures its UEs to accept service from the other network's base stations. The SE applies knowledge of network topology and of services offered. This knowledge is obtained from a sharing database. At the expiration of the sharing agreement, the SE tells the MMEs to deactivate the sharing agreement.

A method for modifying instances of a repeating pattern in an integrated circuit design to correct for perturbations during rendering is described. In the typical embodiment, these corrections are optical proximity corrections that correct for optical effects during the projection of the mask pattern onto the wafer and/or processing effects for example photoresist response and etching effects. The method comprises determining a correction for the repeating pattern based on a first set of tolerances for features of the repeating pattern. Then, the suitability of the corrections is evaluated for instances of the repeating pattern in the integrated circuit design based on a second set of tolerances, which is different from the first set of tolerances. This can be used to preserve much of the hierarchy of the layout data in the corrected, or lithography, data. This can be achieved during the OPC process, thus avoiding the post OPC compaction. It can further take advantage of the fact that, for a given physical layer of a chip for example, different portions of the representing design polygons typically have different requirements on pattern fidelity on the wafer while perturbations may vary as a function of field position. By applying knowledge of the feature tolerances, and allowing design corrections only when tolerances are not met, the data explosion that occurs when moving from layout to lithography data can be contained without sacrificing accuracy.

The invention discloses an SAR ship target detection method based on network pruning and knowledge distillation, and belongs to the technical field of radar remote sensing image application. Accordingto the technical scheme, firstly, a backbone network structure suitable for SAR multi-scale and large-length-width-ratio targets is designed, and then channel pruning is conducted on the network to generate a compact model. In addition, a knowledge distillation strategy is applied to make up performance degradation caused by network pruning. Different from full-feature simulation in a common detection model distillation method, the method has the advantages that the mutual relation between feature maps of different levels is used as migration knowledge in distillation, and a novel attention mechanism is designed to enhance target-related features, so that the distilled features have higher representation capability. By adopting the detector constructed by the invention, the model size of2.8 M, the reasoning speed of more than 200fps and lower calculation cost can be realized, and the detection precision is also improved to a certain extent.

The invention relates to the technical field of knowledge reasoning, in particular to a method and system for generating a knowledge graph by applying knowledge reasoning, comprising six steps, a webmanagement platform, a business platform database DB, an Engine platform, a Hive data source, a TmpDB temporary data source and a DataSource data source. The method comprises the steps that 1, a custom algorithm program is uploaded through the web management platform, and parameters are configured; and 2, an inference scene is configured. According to the method and the system, the problem that the whole process from knowledge reasoning to knowledge graph generation lacks a customizable reasoning algorithm is solved; visual configuration, automated ontology construction, a visual knowledge graph problem is synchronously generated; a user can configure a self-defined algorithm and an inference scene through a visual page, and then inference is conducted on knowledge through a self-developedworkflow management system by means of airflow and spark; and finally ontology construction of the knowledge graph and visual graph display are achieved through a self-developed ETL tool.

A system and method for the automation of architectural changes to a system level design process. Various embodiments of the process include generating product definition templates for each node in a product model, then applying knowledge-based engineering processes and tools in conjunction with interpart modeling functionality to automate the architectural change using the product definition templates.

The invention provides a teaching system based on a cognitive mechanism. The teaching system comprises a user supporting module and a teaching resource management module, wherein the user supporting module is used for carrying out authentication on users and managing education resources according to user instructions after the user authentication, the teaching resource management module is used for searching for and storing teaching resources and displaying the teaching resources according to user instructions, and the teaching resources comprise knowledge logic structure diagrams and thinking form annotation diagrams of chapters and sections of teaching courses. The education resources are managed according to the user instructions after the user authentication by setting the user supporting module, the users can be identified quickly and accurately, and corresponding operations can be carried out according to the user instructions; through the teaching resource management module, the teaching resources comprise the knowledge logic structure diagrams and the thinking form annotation diagrams of the chapters and the sections of the teaching courses are collected and stored, the innovation capability of students can be improved, the students can acquire, remember and apply knowledge quickly and firmly, and thus teaching quality can be improved.

The invention discloses a knowledge point mining method and device. The method includes: mining non-real-time concerns from non-real-time user interaction voice data; mining real-time concerns from real-time user interaction data; generating underlying application knowledge according to the non-real-time concerns and the real-time concerns. The method has the advantages that by mining the concernsfrom the non-real-time and real-time user interaction data, knowledge point mining of the user interaction data under various modes is achieved, and the generated underlying application knowledge iscomprehensive and accurate.

The invention belongs to the technical field of visual identification, and particularly provides a smoke detection method. The invention aims to solve the problems that in the prior art, smoke detection labeling difficulty is large, and many computing resources are occupied. In order to achieve the purpose, the smoke detection method comprises the following steps: marking a small batch of smoke image samples, training a first detection model, detecting the large batch of smoke image samples by using the first detection model to generate pseudo labels, and enabling the first detection model to apply the labeled small-batch smoke image samples and the large-batch smoke image samples with the pseudo labels, and guiding a second detection model to train to obtain the smoke detection model by applying a knowledge distillation algorithm. Through the semi-supervised learning and knowledge distillation algorithm, the workload of marking the smoke image samples is reduced, and meanwhile, a teacher network with high accuracy is applied to train a student network, so that the occupation of computing resources can be reduced under the condition of ensuring the accuracy.

The invention relates to a grade protection evaluation data acquisition and analysis method and system based on an offline form, and belongs to the technical field of network security grade protectiondata processing. The method comprises the steps: defining and generating an offline data acquisition form according to the features of a tested system, and manually acquiring and reporting data; verifying, analyzing and cleaning the acquired data based on an application knowledge base, and then nesting a structured data retrieval engine; calculating and analyzing the structured data, and matchingsafety suggestions according to the risk points; and completing report ontology library construction, and calling a base block library to generate a report based on elements such as a report mother set and an attribute index. The data processing method integrating acquisition, analysis, calculation and display is constructed by taking improvement of the equal insurance evaluation operation efficiency, quantification of risk analysis results, reduction of manual judgment intervention and avoidance of distributed data leakage as targets, and informatization support means of front-end and rear-end work of equal insurance evaluation are enriched and standardized.

The invention discloses a contract defect detection method based on an intelligent contract knowledge graph. The contract defect detection method is mainly used for detecting code defects of an intelligent contract written by a Solidiity language on an Ethereum platform. The code defects of the smart contract refer to errors, faults or vulnerabilities in the smart contract, which may cause incorrect or accidental execution of the smart contract. The defect mode of the smart contract is defined by analyzing the defect of the smart contract. For a to-be-detected smart contract, the source code of the smart contract is analyzed based on Solidity grammar, and a smart contract abstract syntax tree is generated; basic knowledge entities and relationships thereof of the smart contract are obtained through an abstract syntax tree, so that a smart contract knowledge graph is constructed, and the smart contract knowledge graph is stored in an RDF triple database; and on the basis of the intelligent contract knowledge graph, reasoning query is performed by applying a knowledge graph query statement, the intelligent contract is matched with a defect mode, and defects and positions of the intelligent contract are detected.

The invention discloses a knowledge extraction method, which comprises the following steps: S1, transforming personal or organizational experience into electronic intermediate knowledge through text input technology, voice input technology or video input technology, and uploading the electronic intermediate knowledge to a server through a network transmission protocol; S2, classifying the intermediate knowledge according to the content structure elements contained in the intermediate knowledge; S3, extracting the core knowledge points in the intermediate knowledge, and establishing the list ofthe knowledge points of the core knowledge points; S4, using the knowledge extraction tool to structure the content structure elements in the list of knowledge points to form the general knowledge, and upload it to the server through the network transmission protocol. The invention is innovative in that: knowledge is classified, and the inherent law and the structural element law among the knowledge are found according to the knowledge, so as to avoid the knowledge bias and the knowledge loss caused by the lack of the structure.

A method for adjusting a training of a specific user is provided. The method includes steps of: (a) a server acquiring identification information of the specific user from a user terminal, and (b) the server performing or supporting another device to perform, by referring to the identification information, (I) a first process of allowing the specific user to train basic knowledge data and basic motion data on a specific subject by providing a pre-training content to the user terminal; (II) a second process of allowing the specific user to train practice knowledge data or practice motion data on the specific subject by providing a practice training content to the user terminal; and (III) a third process of allowing the specific user to train application knowledge data and application motion data required to solve scenario-based problems on the specific subject by providing a scenario-based training content to the user terminal.

The invention discloses a campus safety internet of things perception platform, which includes a self-service entry end, a network communication terminal and a background server, the self-service entry end is internally provided with a sensing platform, the integrated sensing platform comprises an Internet of things device management module, Communication protocol module, virtual antenna sensing module, application function module, business linkage module, virtual drill module, knowledge base management module, AI algorithm module, big data display module and on-line patrol module are characterized in that: the self-service input terminal provides a user with a manual login platform and performs manual operation; The network communication end provides the communication of the self-serviceinput end, and transmits the data to the background server end and the internet. By applying knowledge base, artificial intelligence and software bus technology, the invention improves the full self-detection level and the veracity of early warning of the campus safety internet of things, and builds a highly sensitive situation sensing platform of the campus safety internet of things.

Data can be categorized as facts, information, assumptions and instructions. Activities that generate certain categories of data based on other categories of data by applying knowledge can be categorized as classification, evaluation, interpretation, and implementation. Activities can be driven by a Classification-Assessment-Analysis-Actuation (CARE) control engine. CARE controls and these categorizations can be used to enhance numerous systems, such as diagnostic systems, such as through history keeping, machine learning and automation. Such diagnostic systems may include systems that predict computing system failures based on applying knowledge to system vital signs such as thread or stack segment strength and memory heap usage. These vital signs are facts that can be classified as generating information such as memory leaks, escort effects, or other problems. Classification can involve automatic generation of classes, states, observations, predictions, norms, objects, and processing of sampling intervals of irregular duration.

The invention discloses a ground flight test platform and test method for a coaxial double-rotor unmanned aerial vehicle and relates to the technical field of unmanned aerial vehicle tests. The groundflight test platform comprises a mechanical part and a control part. a sensor is arranged on the mechanical part, and data are collected through stm32. According to data receiving and data processingparts, the data are received by an HC05 Bluetooth module, and operation is conducted through an upper computer. The ground flight test platform comprehensively applies knowledge of the mechanical, sensing and software fields, specific to the flight characteristics of the coaxial double-rotor unmanned aerial vehicle, the mechanical structure of the platform is simplified, and data collection and data processing functions are added. Flight parameters of the unmanned aerial vehicle can be detected in real time while the unmanned aerial vehicle is protected, and later adjustment is facilitated. The sensor is replaced, thus the ground flight test platform can meet the requirements of different types of unmanned aerial vehicles and detection tasks of different measuring ranges, and good applicability is achieved.

The invention relates to the development of descriptive and normative systems based on sensor data, and relates to machine learning tasks such as classification and regression. Any such system development requires participation of different interested parties to obtain features. These typically obtained features are uninterpretable for 1-D sensor signals. Embodiments of the present disclosure provide systems and methods capable of performing signal analysis for feature extraction and interpretation thereof, wherein an input is original signal data. In the original signal data, the origin of the feature is depicted as signal data and mapped to domain/application knowledge. Features are extracted using a deep learning network, and a machine learning (ML) model is implemented for sensor dataanalysis to perform predicted causal relationship analysis. A layer (e.g., the last layer) of the deep network includes automated export features that can be used for ML tasks. Parameter adjustment isperformed according to a set of features recommended by the system to determine the performance of the considered system (or application).