13 results about "Artificial networks" patented technology

The invention discloses a domain model and platform engine docking system and method for a multi-agent calculation experiment system. The method is based on a domain knowledge analysis module, an agent modeling docking module and an environment modeling docking module, and rule evolution modeling comprises the following steps: the domain knowledge analysis module analyzes a domain knowledge document input by a user through a large language model; automatically extracting experimental system boundaries, key roles, environmental elements and behavior logic relationships, and outputting domain knowledge information matched and analyzed with each docking module; the agent modeling module establishes a relationship with a role model in an artificial society based on the analyzed domain knowledge information; the environment modeling docking is based on a standard geographic space and an artificial network establishment relationship; the rule evolution modeling establishes a relationship with an artificial network through a strategy feedback dynamic interaction rule based on individual learning, group emergence and a regulation and control evolution mechanism; the platform engine is used for mapping analysis domain knowledge information to a calculation experiment platform through a standardized interface; according to the method, a complete technical path is provided for efficient construction and simulation analysis of an artificial social system through a structured, standardized and adjustable model construction and docking mechanism.

The application relates to the field of artificial network control, in particular to an artificial network control method, device and system for shielding across a darkroom. By using a fiber link and an internal battery power supply architecture in a fully-closed shielding darkroom, the interference of personnel on a test environment is completely eliminated, monitoring of parameters such as internal temperature and voltage of the integrated darkroom is added, zero-crossing synchronization noiseless operation of phase switching is realized, in addition, user end environment adaptive identification, driving parameter dynamic calibration and equipment life prediction are realized, and the needs of high-precision, unattended automatic phase switching and testing are met.

Various embodiments of the present technology generally relate to robotic devices and artificial intelligence. More specifically, some embodiments relate to an artificial neural network training method that does not require extensive training data or time expenditure. The few-shot training model disclosed herein includes attempting to pick up items and, in response to a failed pick up attempt, transferring and generalizing information to similar regions to improve probability of success in future attempts. In some implementations, the training method is used to robotic device for picking items from a bin and perturbing items in a bin. When no picking strategies with high probability of success exist, the robotic device may perturb the contents of the bin to create new available pick-up points. In some implementations, the device may include one or more computer-vision systems.

This invention relates to the field of artificial power network technology, specifically to a high-voltage artificial power network system and equipment. It includes a DC decoupling network module integrating decoupling capacitors and discharge resistors for power protection; and a DC load network module electrically interconnected with the DC decoupling network module via cables, integrating a multi-stage resistor array, a heat exchanger system, control components, and a cooling subsystem. The cooling subsystem is configured for water-air heat exchange or water-water heat exchange based on the current output stage of the multi-stage resistor array. This invention features efficient heat dissipation and precise temperature control, real-time closed-loop temperature monitoring, and over-temperature alarms for safety. It can promptly observe localized temperature rises and water levels caused by poor cable contact. Connecting to a host computer via a communication module allows for real-time monitoring by the host computer software, and interlocking protection is implemented with the power supply equipment and the tested equipment, simplifying maintenance.

The invention relates to the technical field of power supply radiation testing, in particular to a power supply on-line conduction and radiation integrated testing system and method, and the system comprises a testing case, and a spectrum analyzer, an artificial network and an isolation power supply which are disposed in the testing case. The industrial control computer is in communication connection with the test case, and the industrial control computer is loaded with a system program for controlling a test process and analyzing data; the test unit is connected with the test case through a radio frequency transmission line and comprises a desktop shielding box, a test jig arranged in the desktop shielding box and a radio frequency current probe used for detecting radiation signals; wherein the system program can control the frequency spectrum analyzer to switch working modes so as to carry out conduction disturbance test or radiation disturbance test on a tested power supply arranged on the test fixture. According to the invention, integration and automation of testing are realized.

A device for calculating thermal characteristics of a building structure using an artificial neural network comprise a first artificial neural network which receives first input information including a drawing image as input information and outputs first output information including feature information, in which shape information for each part of the drawing image is extracted, as output information, a second artificial network which receives the first output information and second input information including thermal conductivity information for each part as input information and outputs second output information including temperature distribution image information corresponding to the drawing image as output information and a third artificial neural network which receives the first output information and the second input information as input information and outputs third output information including heat flow distribution image information corresponding to the drawing image as output information.

The application discloses a kind of infectious disease network key node identification method based on graph neural network, comprising the following steps: 1, by traditional SIR infectious disease model on the generated network simulates virus transmission process, constructs training data label;2, select the highest score of Kendall coefficient and the score obtained in step 1 is combined with feature, the selected network features are input into the attention model with multiple graph attention layers are introduced, in a variety of artificial networks are trained, to obtain the importance score ranking of each node more suitable;3, consider the "rich club" effect, the obtained ranking sequence is selected by the node selection algorithm based on distance 2-hop, select the node set that the influence range is further expanded.The application provides a new idea and tool for key node identification in complex network, and provides a scientific basis for network optimization and risk prevention.

The invention relates to a system (110) comprising an input projection layer (111), at least one second digital artificial neural network (50), and at least one semiconductor chip (40) having an analogous artificial neural network (10) being mapped from a trained first digital artificial network (17) wherein the analogous electrical circuit comprises at least one input (20) being configured to receive at least one input signal and at least one output (26) being configured to provide at least one analogous output signal, wherein the analogous artificial neural network (10) electrically connects the input (20) to the output (26), wherein the semiconductor chip (40) is recursively coupled to the second digital artificial neural network (50) such that the output of the semiconductor chip (40) is provided to at least one hidden layer of the second digital artificial neural network (50) and wherein the input (20) of the semiconductor chip (40) and an input layer of the second digital artificial network (50) are electrically connected to the input projection layer (111). The invention provides a fast processing artificial neural network (10) that operates faster than the second digital artificial neural network. Thus, the analogous artificial neural network provides output to influence the operation of the second digital artificial neural network before the second digital artificial neural network provides output.

The application provides an electric drive bench background noise verification method for an electromagnetic compatibility darkroom. An electric drive system is arranged in the electromagnetic compatibility darkroom. The electric drive system comprises a motor. A motor shaft is connected with a dynamometer shaft. A motor controller is connected with a battery simulator through a high-voltage artificial network. A set rotating speed is given to the dynamometer. The motor is dragged to rotate at a stable rotating speed. The battery simulator is adjusted to a load mode to absorb power supply of the electric drive system. After the working condition is stable, an antenna for detecting noise is placed at a position with a human-set distance from a table edge. The result of the antenna receiver is compared with a reference value. If the result is lower than a required value, the background noise requirement is met. If the result is higher than the required value, the background noise requirement is not met. The electric drive bench background noise verification method for the electromagnetic compatibility darkroom combines the electric drive system with the dynamometer and the battery simulator. The electric drive system is under a load condition. Background noise testing is performed. The result is more real, accurate and effective.

The application relates to the field of artificial network control, in particular to an artificial network control method, device and system for shielding across a darkroom. By using a fiber link and an internal battery power supply architecture in a fully-closed shielding darkroom, the interference of personnel on a test environment is completely eliminated, monitoring of parameters such as internal temperature and voltage of the integrated darkroom is added, zero-crossing synchronization noiseless operation of phase switching is realized, in addition, user end environment adaptive identification, driving parameter dynamic calibration and equipment life prediction are realized, and the needs of high-precision, unattended automatic phase switching and testing are met.

The invention relates to a matched rice quality prediction method based on an artificial neural network. The method comprises the following steps: step 1, preparing raw materials and a matched rice sample; step 2, based on the raw materials and the matched rice samples obtained in the step 1, measuring the taste value, TPA characteristics, physicochemical indexes and rice adhesion layer index values of the matched rice; step 3, establishing a rice taste value and TPA prediction model by using an artificial network model; and step 4, based on the physicochemical indexes of the raw material rice and the rice adhesion layer index value obtained in the step 2, calculating the physicochemical indexes of the matched rice samples and the rice adhesion layer index value, and based on the rice taste value obtained in the step 3 and the TPA prediction model, predicting the quality of the matched rice so as to screen out the matched rice sample with the optimal taste value. According to the invention, synchronous prediction of rice taste value, hardness, viscosity and elasticity indexes of the compound rice is realized.

The invention relates to the technical field of high-voltage safety testing of new energy automobiles, in particular to a broadband high-precision artificial network impedance circuit which is used for verifying voltage transient tolerance and impedance characteristics of high-voltage parts of new energy automobiles and comprises a main circuit module and a safety isolation module. The main circuit module comprises a low-inductance resistor, an adjustable inductor and a high-frequency compensation capacitor which are connected in sequence, the low-inductance resistor is used for current sampling, the inductance value of the adjustable inductor can be adjusted to cover the frequency band of 10Hz-10MHz, the high-frequency compensation capacitor is used for suppressing the resonance peak of the circuit, and the high-frequency compensation capacitor is used for compensating the resonance peak of the circuit. The safety isolation module comprises a voltage sensor, a Hall current sensor, a shielding box and a grounding plane which are subjected to optical coupling isolation, and common-mode interference between the shielding box and the grounding plane is suppressed through a Y capacitor. Through voltage fluctuation testing and impedance characteristic testing, high-precision testing of high-voltage parts is achieved, the high-frequency-band impedance error is small, and the requirements for voltage fluctuation, impedance testing and safety isolation are met.

The application relates to the technical field of vehicle testing, and discloses an electromagnetic testing system and method for a vehicle-mounted electric drive system. The system comprises a metal platform placed in a darkroom, an electric drive system placed on the metal platform, the electric drive system connected to a dynamometer through a first transmission shaft, a first transmission wheel on the first transmission shaft and a second transmission wheel on a second transmission shaft connected through a belt transmission, and a power generation input interface of the electric drive system connected to the second transmission shaft. A weak current power supply port of the electric drive system is connected to a weak current artificial network through a weak current line and then connected to a weak current power source. A strong current power source is arranged outside the darkroom, a strong current power supply port of the electric drive system is connected to a strong current artificial network through a strong current line and then connected to the strong current power source by penetrating the wall of the darkroom. A collection terminal is in communication connection with an upper computer outside the darkroom, and the upper computer is in communication connection with the electric drive system through a communication module. The application provides accurate data for human body electromagnetic protection design of the electric drive system under different working conditions and comprehensively guarantees the health and safety of drivers and passengers.