21 results about "Neuron structure" patented technology

The invention discloses a robot virtual-real cooperative training decision optimization system and method based on digital twinning, and the method comprises the following steps: collecting state data and disturbance data of an entity robot in a real environment, and carrying out the preprocessing of the data to generate standardized input; joint coding and state perturbation mapping are carried out on the standardized data, and a feature vector sequence embedded in a hyperspherical manifold space is generated; inputting to a virtual twin control body based on a hypersurface neural element structure, executing disturbance direction sensitive activation, and outputting an activated state vector sequence; virtual and entity control action prediction sequences are generated respectively, an embedded space difference vector is calculated, and control body parameters are updated based on a consistency optimization criterion; after convergence, the control body executes reasoning to generate a target control action sequence, the entity robot is driven to complete action execution, and control strategy optimization is achieved. According to the method, high-precision migration and rapid convergence of a robot control strategy are realized, and the execution stability in a complex disturbance environment is improved.

The disclosure provides a multi-modal neuron structure, comprising a reference voltage generation module and a pulse generation module, the reference voltage generation module comprises a first reference voltage generation module and a second reference voltage generation module, the first reference voltage generation module generates a first reference voltage, and the second reference voltage generation module generates a second reference voltage; the pulse generation module is connected with the reference voltage generation module, an input end of the pulse generation module is used for receiving a frequency-coded pulse sequence or a time-coded pulse sequence, and the pulse generation module comprises a first charge packet counter sub-circuit, a second charge packet counter sub-circuit and a mode signal input circuit. The disclosure can receive a frequency-coded pulse sequence or a time-coded pulse sequence, has a simple structure, occupies a small area, and is suitable for various forms of pulse sensors.

The invention relates to a method for marking neurons and application of the neurons to unicellular reconstruction of the neurons. The method comprises the following steps: expressing a proximity marker enzyme with enhanced solubility in a target neuron, and providing biotin to the target neuron so as to form biotinylated protein in the neuron; carrying out integral sample dyeing on the biotinylated protein by using monovalent streptavidin coupled with a detectable marker; imaging the dyed sample to obtain a neuron structure image; and reconstructing the morphological structure of the single neuron based on the image data.

The invention relates to the technical field of motor control, and discloses an electrical transmission system parameter self-tuning system based on a neural network, and the system comprises a state monitoring and gating module which is used for activating the system when the high-frequency energy characteristic of a torque instruction exceeds a threshold value; the homomorphic mapping and reference reconstruction module is used for constructing a homomorphic digital filter to convert a torque instruction into a homomorphic instruction aligned with a feedback speed time sequence; the momentum residual calculation module is used for calculating the difference value between the theoretical impulse and the actual momentum increment based on the momentum theorem; the morphological decoupling and reasoning module is used for calculating morphological similarity between the residual vector and the reference vector by using the neuron structure so as to output confidence; and the parameter updating module is used for adjusting the adaptive updating gain according to the confidence coefficient and iteratively correcting the inertia parameter. According to the method, through signal homomorphic reconstruction and waveform form reasoning, the influence of loop delay and load disturbance is effectively inhibited, and the accuracy and robustness of parameter setting are improved.

This invention provides a research model and construction method for calcium ion imaging of ASH neurons in nematodes, belonging to the field of biological model construction technology. The construction method includes the following steps: nematode culture; construction of a transgenic probe plasmid: obtaining the ASH neuron-specific promoter sra-6 using PCR technology, and connecting HindIII and BamHI restriction sites to both ends of the promoter; amplifying YC3.60 using PCR technology; then inserting the promoter sequence into the pPD95.75 plasmid, and then inserting YC3.60 after the promoter. Transgenic microinjection of nematodes is used, followed by screening to obtain nematodes with ASH neurons carrying green fluorescent protein. In this application, YC3.60 is transferred into nematodes via transgenic means. Using a promoter specifically expressed in ASH neurons to link the gene of this fluorescent probe protein into the nematode, the fluorescent probe can be specifically expressed in the neuron. The constructed nematode model can be used for neuronal structure and functional imaging studies, as well as neuropharmacological efficacy analysis.

This invention relates to nutritional compositions and their uses. Specifically, this invention relates to a nutritional composition comprising: (1) component A consisting of docosahexaenoic acid and arachidonic acid; (2) component B consisting of medium- and long-chain fatty acid triglycerides and 1,3-dioleoyl-2-palmitoyl triglycerides; and (3) component C consisting of choline and taurine. This invention also relates to the use of this nutritional composition for non-therapeutic purposes, including: promoting neurotransmitter expression; improving neuronal structure and function; and improving anxiety. This invention further relates to the use of this nutritional composition in the preparation of products for non-therapeutic purposes, including improving brain development and / or memory and / or improving anxiety. The nutritional composition of this invention can significantly improve brain development and / or memory and / or improve anxiety, which is beneficial for meeting current market demands.

Disclosed herein are compounds, compositions, and methods for promoting neuronal growth and / or improving neuronal structure with the compounds and compositions disclosed herein. Also described are methods of treating diseases or disorders that are mediated by the loss of synaptic connectivity and / or plasticity, such as neurological diseases and disorders, with fused pyrrolidine psychoplastogens.

This invention discloses an optical neuron structure based on an on-chip microring resonator, comprising, from front to back, a weighting unit, a summing unit, and a nonlinear unit. The weighting unit employs a wavelength division multiplexing (WDM) series microring resonator structure to apply weights to multiple input optical signals. The summing unit includes a photodetector and a preamplifier, used to sum the multiple weighted optical signals and drive the microring resonator of the nonlinear unit. The nonlinear unit employs a microring resonator structure to perform nonlinear output processing on the summed optical signals, primarily utilizing the nonlinear transmission characteristics of the microring resonator. This invention features high integration, and the proposed optical neuron exhibits strong scalability. Furthermore, the on-chip integration process is compatible with CMOS, enabling large-scale fabrication and facilitating the miniaturization, high integration, and high scalability of the optical neuron.

The application provides a phishing email recognition model and method based on multi-level multi-features and deep learning, the recognition model comprising a multi-channel input layer, a plurality of sub-channel input layers, and the total number of the sub-channel input layers being configured to correspond to the total number of input multi-level feature types; a multi-level embedding layer comprising a word-level embedding layer and a character-level embedding layer, the total number of the sub-embedding layers being configured to the total number of types of first features, and the first features being input to map from high dimension to low dimension to output third features; a BILSTM layer being configured to correspond to the sub-embedding layer and connected to the sub-embedding layer, the third features being input to splice and output fourth features; a feature joint layer being configured to input the second features and the fourth features, fuse and output joint features; and a full connection layer being configured to have a multi-layer neuron structure, the joint features being processed through the multi-layer neuron structure and output through an output layer structure, and the application has high efficiency and good robustness in recognizing phishing emails.

A system simulating a decisional process in a mammal brain about characteristics of motions related to body gestures of a visually observed body through a simulated visual path is provided. The system includes an interface toward simulated neuronal structures, the interface at least converting luminous information of the observed body to an optic flow data stream conveying information related to the visually observed body and that can be processed in the simulated neuronal structures, the system being a feed-forward system and comprising hierarchically from the visual observation to the decision: the simulated visual path and its interface, a simulated local motion direction detection neuronal structure for the detection of motion directions with receptive fields, a simulated opponent motions detection neuronal structure, a simulated complex patterns detection neuronal structure, and a simulated motion pattern detection neuronal structure.

This invention relates to a method and apparatus for converting an attractor neural network into a digital spiking neural network. The method includes: acquiring the neuronal structure of the attractor neural network; establishing a one-to-one correspondence between each neuron in the neuronal structure of the attractor neural network and a large neuron in the digital spiking neural network, each large neuron including a receptor surface state neuron and a clock excitation neuron; the connection relationship of the digital spiking neural network is as follows: the clock excitation neuron in each large neuron is connected to the receptor surface state neurons in other large neurons (excluding itself), and the receptor surface state neurons in each large neuron are connected to the receptor surface state neurons in other large neurons (excluding itself). Compared with the prior art, this invention has the characteristics of high stability, strong anti-interference, and good real-time performance of parallel data compared with traditional serial input data.

The application discloses a channel estimation method, device, equipment, storage medium and program product, relates to the wireless technical field, and the method comprises the following steps: preprocessing channel state information of a frequency domain and a time domain of a communication network to obtain preliminary channel estimation data; optimizing the preliminary channel estimation data by a preset spike neural network adopting a multilayer dynamic neuron structure to generate optimized channel estimation data; extracting time-frequency domain state information of a channel based on joint time-frequency domain data converted by the optimized channel estimation data; and performing channel characteristic analysis on the time-frequency domain state information to generate a channel estimation result. Since the application improves the spike neural network by adopting the multilayer dynamic neuron structure in advance, the channel characteristic analysis is performed on the time-frequency domain state information corresponding to the generated optimized channel estimation data, and a high-precision channel estimation result can be obtained, thereby effectively improving the accuracy and robustness of channel estimation.

The invention belongs to the technical field of reservoir water temperature prediction and regulation and control, and particularly discloses an upstream discharged water temperature prediction and flexible water retaining curtain wall regulation and control method and system. According to the method, a physical mechanism of an LSTM neuron structure is reconstructed, a water body stratification stability mechanism and a vertical thermal diffusion mechanism are deeply fused in a deep learning network, an improved LSTM model can well predict a water temperature stratification phenomenon of upstream incoming water, and high-precision prediction of vertical water temperature distribution of a reservoir under a complex meteorological condition is realized. A water temperature mixing model calculation formula is created for the first time, and the problem that the discharged water temperature requirement is difficult to calculate is solved. The vertical water temperature predicted by the improved LSTM is combined with the discharged water temperature inverted by the downstream ecological target, and the optimal water taking elevation is compared and determined, so that accurate lifting regulation and control of the flexible water retaining curtain wall are guided, the ecological water temperature requirement of a downstream river channel is strictly guaranteed, meanwhile, water resource waste is effectively avoided, and the power generation benefit and the flux utilization rate of a reservoir are maximized.

The invention discloses a multi-link mechanism optimization method and system based on a neural network and adaptive SQP, and the method comprises the steps: constructing a training sample with the length of each link as the input and the blanking speed, maximum speed and maximum acceleration of a sliding block as the output, carrying out the modeling through a feedforward neural network, optimizing a weight parameter through a multi-layer neuron structure and a gradient descent algorithm, and carrying out the optimization of a multi-link mechanism. Constructing a neural network model; performing an input variable disturbance experiment on the neural network model, extracting the sensitivity of each design variable to an output performance index, and forming a sensitivity matrix and a parameter response trend chart; according to the method, the blanking speed, the maximum speed and the maximum acceleration of a sliding block are comprehensively considered, a weighted multi-objective function is constructed with minimization as a target, a neural network model is used for replacing real simulation to carry out performance prediction on a design variable combination, and a self-adaptive SQP algorithm is used for solving for constraint processing and parameter discretization in the multi-objective optimization process. The neural network modeling and the adaptive SQP algorithm are combined, and the method is accurate and efficient.

The invention discloses a special transformer user voltage monitoring method and system based on a binary collaborative optimization voltage fluctuation model, and relates to the technical field of power system voltage monitoring. The neuron adaptive optimization algorithm can dynamically adjust the neuron structure of the model to match the time sequence complexity of the voltage data, and the prediction precision of the optimal voltage fluctuation model on the voltage fluctuation is greatly improved through the synergistic effect of the neuron adaptive optimization algorithm and the neuron adaptive optimization algorithm, so that the timeliness and accuracy of voltage out-of-limit abnormity decision are ensured. The technical problem that the prediction precision of a voltage fluctuation prediction model in the existing special transformer user voltage monitoring technology is insufficient is effectively solved, and safe and stable operation of a power grid and special transformer user equipment is guaranteed.

The application provides a modular decomposition method for a Transformer neural network model, modular training is performed on a randomly initialized model, and structured decomposition is implemented by applying a mask after the modular training, so that flexible on-demand reuse is realized; specifically comprising the following steps: S1, modular training; given a neural network model, firstly, randomly initialize all parameters of the model; a neuron recognizer is included, which is used to identify neurons related to a specific function; S2, structured decomposition; the modular training model is subjected to modular decomposition; S3, on-demand reuse; on-demand reuse is realized by structurally removing neurons, and memory and computing overheads are minimized. The application can realize structured decomposition of a Transformer structure model and flexible on-demand reuse of functions, and has high expandability.

Disclosed herein are compounds, compositions, and methods for promoting neuronal growth and / or improving neuronal structure with the compounds and compositions disclosed herein. Also described are methods of treating diseases or disorders that are mediated by the loss of synaptic connectivity and / or plasticity, such as neurological diseases and disorders, with non-hallucinogenic psychoplastogens.

The invention relates to the technical field of crop growth, in particular to a simulation calculation method based on a crop physiological ecological model, which comprises the following steps: collecting crop wave frequency response data obtained by a wave frequency sensor, the wave frequency response data comprising a reflection signal, a transmission signal or a scattering signal of a crop tissue under the action of external wave frequency; and constructing a crop physiological ecological model based on the neuron structure, wherein the crop physiological ecological model comprises a plurality of physiological state units used for representing a photosynthetic process, a transpiration process, a moisture regulation process and a hormone regulation process, and a connection weight used for establishing a mapping relationship among the environmental data, the wave frequency response data and the physiological state units. According to the method, the crop wave frequency response data are collected, and the crop physiological ecological model is constructed based on the neuron structure, so that the problem that crop management is inaccurate due to the fact that the crop physiological state in traditional agriculture is invisible and not quantified is solved.

This application provides a communication method and a communication device, applicable to the field of communication. In the technical solution of this application, when the access network device sends the model to the terminal by sending the weights in the neuron structure of the model, the weights are first quantized before sending the quantized weights, thereby reducing the transmission overhead of the model.

The invention discloses application of a snakegourd peel injection in preparation of a medicine for treating spinal cord injury, and relates to the technical field of biological medicines. SCI mouse model research finds that TPI can promote motor function recovery after SCI, relieve tissue damage, protect neuron structures, inhibit neuroinflammation, promote polarization of microglial cells to anti-inflammatory phenotypes, relieve oxidative stress after spinal cord injury, inhibit neuronal apoptosis and promote axonal regeneration, and can be used for preparing the medicine for treating the spinal cord injury. The potential action mechanism of the TPI is disclosed through network pharmacology and molecular docking, and the high consistency of an in-vivo experiment result and an in-vitro prediction result enhances the rationality of taking the TPI or an active component thereof as a candidate drug for SCI treatment.

The application discloses an image classification method based on a multi-stage complementary pulse neural network model. The method comprises the following steps: acquiring an image classification data set; constructing a multi-stage complementary pulse neural network model, which comprises the following steps: using a traditional LIF neuron structure to construct an SNN network model; reconstructing the LIF neuron structure in the SNN network model and introducing a multi-stage activation design, wherein the multi-stage activation design uses a plurality of small neuron structures to construct a new multi-stage neuron structure; after the multi-stage activation design is introduced, a complementary potential is introduced, a complementary activation design is realized, and a multi-stage complementary neuron structure is obtained; and the multi-stage complementary pulse neural network model is trained by using the image classification data set, and the trained multi-stage complementary pulse neural network model is used for image classification. The application improves the accuracy of image classification.