37 results about "Neuronal activation" patented technology

The invention discloses a hardware acceleration system and method supporting a pulse-gated recurrent neural network, and relates to the technical field of neural network hardware acceleration, and the system comprises a PC terminal and an accelerator. The accelerator comprises a UART (Universal Asynchronous Receiver / Transmitter) module, a data composer, a storage module, a controller module, a calculation module and a gating calculation core, the PC terminal is connected with the UART module; the gating calculation core is connected with the data composer; and the controller module and the storage module are respectively connected with the UART module, the data composer, the calculation module and the gating calculation core. Based on the above scheme, different coding modes can be selected at a PC terminal to adapt to different data types, and a pulse neuron activation function and a logic gate are utilized to perform calculation heterogeneous design in a calculation module and a gating calculation core, so that a high-performance configurable hardware acceleration calculation architecture for the pulse gating recurrent neural network is realized.

There are provided systems and methods for data privacy protection and removal for artificial intelligence model training and deployment. An online transaction processor or other service provider may provide computing services and platforms to entities, which may include use of machine learning (ML) models including large language models (LLMs). To comply with data privacy protections and copyright enforcement, a system may provide unlearning of content from ML models. The system may receive a request to unlearn a content and, after verifying the request is valid, identify the content used for during training of or inferencing by an ML model. The system may then map the content to concepts and correlate those concepts with ML model outputs using projections in a vector space. Based on the mapped concepts and outputs, neuron activation of the ML model may be analyzed to identify a negation vector and perform selective parameter dampening.

The invention discloses a road target detection method and device based on multi-sensor fusion, and belongs to the technical field of target detection, and the method comprises the steps: collecting the multi-modal sensor data of a target road and the position information of each sensor; based on the position information of each sensor and the multi-modal sensor data, analyzing a space-data communication relationship among the sensors, and constructing a sensing adjacency matrix; extracting spatio-temporal characteristics of the multi-modal sensor data based on the sensing adjacency matrix, generating a spatio-temporal characteristic set, and obtaining a neuron activation value of each calculation path in real time; performing redundancy analysis on each calculation path according to the neuron activation value of each calculation path, and screening out a reserved path; performing redundancy removal processing on the space-time feature set based on each reserved path to obtain a sparse space-time feature set; and performing data fusion on the sparse space-time feature set to obtain a target detection result of the target road. By implementing the method and the device, the problem of low target detection accuracy in the prior art can be solved.

The invention discloses an optimal path planning method of a pulse coupling neural network based on double constraints. The method comprises the following steps: mapping a path planning environment to a DC-PCNN network; a DC-PCNN neural network model is constructed, and all neurons are initialized; activating a target neuron, and recording the current neuron as a father node; calculating an exponential decay function of the torque deviation, and multiplying the calculated value as a penalty factor by an update item of the internal activity item; calculating a gravitational function value of the flow field constraint, and using the calculated value to update a current neuron dynamic threshold value; comparing the internal activity item with a dynamic threshold; when gt; if yes, activating the neuron, and recording the current neuron as a father node; repeating the steps S4-S6 until the initial neuron is activated; and backtracking all activated nodes, and planning an optimal path. According to the method, the search efficiency is remarkably improved while the path optimality is ensured.

The application discloses a video recommendation historical bias mitigation method based on bias attention, which comprises the following steps: (1) data preprocessing; (2) searching for bias neurons; (3) calculating bias masks; (4) calculating bias attention masks; (5) outputting bias attention masks; (6) repeating steps (2), (3), (4) and (5) for each layer until all layers are processed by the bias operation and the output layer is ended. The application avoids the problem of high calculation complexity caused by massive data; the bias attention suppresses the activation output of the bias neurons without changing the normal neuron activation value, and can effectively offset the bias behavior caused by the bias neurons.

The lithium ion battery SOC prediction method based on the CEEMDAN and the optimized BiLSTM comprises the following steps: decomposing preprocessed operation data by adopting adaptive noise complete ensemble empirical mode decomposition CEEMDAN to obtain a plurality of intrinsic mode functions (IMF) and a residual component; constructing a BiLSTM model, wherein the model comprises a forward LSTM unit, a backward LSTM unit and an output layer; a dropout optimization algorithm is introduced into the BiLSTM model, and a combination of a two-parameter sigmoid activation function and a softsign activation function is adopted as a neuron activation function; and inputting the intrinsic mode function IMF and the residual component into the optimized BiLSTM model, optimizing model parameters through training, and predicting the state of charge SOC of the lithium ion battery by using the trained model. Aiming at the problems of weak data nonlinear processing capability, model overfitting, gradient dispersion and the like in the existing SOC estimation method, the invention provides a lithium ion battery SOC prediction method based on CEEMDAN and BiLSTM optimization, so as to improve the SOC prediction precision and model stability.

The application discloses a federal fair framework training method and device based on neuron activation difference. In the process of federal learning, in addition to requiring the client to upload the local model update parameter, the server also requires the client to upload the activation frequency of neurons, and the server evaluates the contribution degree of each client by calculating the sparsity of the neuron activation frequency of the client, so as to issue different quality global models to various clients.

Disclosed herein are inhibitors of aminocarboxymuconate semialdehyde decarboxylase (ACMSD), pharmaceutical compositions and methods for using these compounds to treat or prevent one or more ACMSD-related diseases such as, for example, muscle structure disorders, neuronal activation disorders, muscle fatigue disorders, muscle mass disorders, beta oxidation diseases, metabolic diseases, cancer, vascular diseases, ocular vascular diseases, muscular eye diseases, liver diseases, inflammatory diseases or renal diseases.

The application discloses a multi-style text migration method based on a shared neuron activation modulation network, and belongs to the technical field of natural language processing, large language model neuron regulation and text style migration. The method first constructs a style dimension and neuron mapping relationship, and divides exclusive neurons and shared neurons; directional activation modulation is performed on the exclusive neurons, and a embedded shared neuron activation modulation network (SNAR) is constructed; through style interaction embedding, joint coding, context modeling and content cross attention decoding, the shared neuron activation distribution is dynamically reorganized, and multi-style representation conflicts are eliminated; finally, the activation values after modulation of the two types of neurons are fused, and multi-style text is generated through model reasoning. The application can efficiently realize cross-dimension multi-style collaborative migration without fine-tuning the large language model, solve the style conflict, semantic deviation and activation competition problems of existing methods, and is suitable for human-computer interaction, content creation and other scenes.

The invention discloses a method for measuring defective neurons of a convolutional neural network model of a migrated brain neural network, and belongs to the technical field of intelligent measurement and testing. According to the method, on the basis of a cognitive activation hypothesis theory, accurate measurement of internal defects of a model is realized by constructing a'traceability chain 'reflecting a defect propagation path and positioning'suspicious neurons' based on differential statistics; the method specifically comprises the following steps: detecting abnormal neurons, and generating and combining high-weight paths to construct a tracing chain; meanwhile, key suspicious neurons are counted and positioned by comparing neuron activation differences under similar input. According to the method, weak links in a model decision-making mechanism can be accurately revealed, the efficiency and interpretability of defect positioning are remarkably improved, and an effective technical support is provided for enhancing the robustness of the model in a safety-critical scene.

Disclosed herein are aminocarboxymuconate-semialdehyde decarboxylase (ACMSD) inhibitors, pharmaceutical compositions, and methods of using these compounds to treat or prevent one or more ACMSD-related diseases, such as a muscle structural disorder, a neuronal activation disorder, a muscle fatigue disorder, a muscle mass abnormality disorder, a beta-oxidation disease, a metabolic disease, a cancer, a vascular disease, an ocular vascular disease, an ocular muscle disease, a liver disease, an inflammatory disease, or a kidney disease.

A method for generating a device-specific artificial intelligence model includes obtaining, by a hardware device, unique information of a specific device, generating, by the hardware device, a neuron activation matrix based on the unique information, setting, by the hardware device, a degree of activation for neurons by applying the neuron activation matrix to neurons of a neural network model, and performing, by the hardware device, fine-tuning on the neurons for which the degree of activation has been set using a positive dataset.

The invention belongs to the field of artificial intelligence, discloses a deep neural network robustness enhancement method for fragile neuron adversarial training, and aims to solve the problem that the natural precision is severely reduced in an existing adversarial training method. According to the method, the concept of a positive value pushing sample is introduced, and generation of the positive value pushing sample with certain aggressiveness is guided by accurately positioning and activating fragile neurons of the model. A target optimization layer is identified by using neuron activation difference indexes of a clean sample and an adversarial sample, and a neuron activation optimization function is designed. And iteratively generating a positive value pushing sample through a gradient optimization method, so that the positive value pushing sample explores an area which is not fully covered in the DNN in a feature space. And the difference between the generated sample and the clean sample in the data manifold is smaller than that of the traditional adversarial sample. Experiments prove that the method keeps the natural precision of the model to the greatest extent while improving the robustness of the model, and is suitable for practical application with high requirements on the natural precision.

Embodiments of the present application provide a neuron activation prediction method and an electronic device. The neuron activation prediction method comprises: performing local calculation based on input features of a current layer of a neural network to generate a local vector for neuron activation prediction; performing activation prediction and screening operation on a plurality of candidate neurons in the current layer of the neural network based on the local vector to obtain a target neuron set and a sparse execution index table corresponding to the target neuron set; performing multiplication and addition calculation and activation function processing on target neurons in the target neuron set based on the sparse execution index table to obtain a sparse activation output result of the current layer of the neural network; determining prediction error information based on the sparse activation output result and an actual activation result, and determining a neuron screening strategy for a next layer of the neural network based on the prediction error information, which can reduce the memory bandwidth demand and energy consumption overhead in the inference process of the neural network.

The invention provides a traffic logistics-oriented large model knowledge base construction method and system, and the method comprises the steps: carrying out the semantic relation mining based on a standard term set in the traffic logistics field, generating a structured term graph, constructing a template knowledge probe set, inputting a pre-training large language model, and carrying out the directional activation detection operation, thereby obtaining an implicit knowledge neuron cluster, the method comprises the following steps of: carrying out space-time correlation analysis on an activation response mode of the neural network, extracting distribution characteristics of neuron activation intensity and a dependency relationship among the characteristics, mapping an implicit knowledge neuron cluster into a semantic traffic logistics field knowledge unit, carrying out hierarchical organization and semantic linking according to a logic correlation degree and a functional attribute classification result, and carrying out hierarchical classification on the neural network. Generating a traffic logistics field structured knowledge network; and calling a standard traffic logistics business problem set to perform knowledge utility verification, and dynamically optimizing and adjusting knowledge units and association relationships to obtain a large model knowledge base, thereby providing high-quality knowledge support for intelligent application in the field of traffic logistics.

The invention discloses an artificial intelligence model optimization method based on brain-like neural network neuron mapping, and the method comprises the steps: obtaining a training data set comprising a plurality of data subsets which correspond to different types of artificial intelligence tasks; inputting the data subsets into an artificial intelligence model for forward propagation, carrying out neuron activation state discrimination, and determining a mapping relation between each data subset and the corresponding activation neuron set; generating a subtask mask matrix of an artificial intelligence model network weight for each data subset; and determining a corresponding sub-task mask matrix according to the type of the to-be-processed artificial intelligence task, and reconstructing the forward propagation path of the artificial intelligence model, so that only the activation neuron set corresponding to the sub-task mask matrix participates in the calculation process of the model when the to-be-processed artificial intelligence task is executed. According to the method, the calculation cost of an artificial intelligence model is greatly reduced, and new-generation brain-like artificial intelligence with anthropomorphic characteristics is constructed.

The application provides a haptic object recognition method based on a pulse graph neural network, and the algorithm comprises the following steps: obtaining a haptic graph, constructing an M-tree haptic graph or a Z-tree haptic graph based on the haptic graph; establishing an object recognition model, training the object recognition model by using the M-tree haptic graph or the Z-tree haptic graph, and the object recognition model comprising an LIF pulse neuron, a topological adaptive graph convolution layer, a full connection layer and a final voting layer; optimizing the object recognition model by using a Gaussian prior distribution loss and an object recognition model reverse propagation training loss; and judging the category of the haptic object by using the optimized object recognition model. The Gaussian prior distribution loss and the object recognition model reverse propagation training loss are weighted, so that the accuracy and stability of the object recognition model for object recognition are improved. Seven approximate functions are used to approximate the partial derivative of the LIF pulse neuron activation function, so that the accuracy of the algorithm is improved.

The invention provides an integrated photon neural mimicry computing chip based on mode division multiplexing, which comprises a multi-mode photon synapse array which comprises N * N multi-mode photon synapses arranged in an array mode, two adjacent multi-mode photon synapses arranged along a row, a first multi-mode photon synapse and a second multi-mode photon synapse arranged along a column, single-mode waveguides of the multi-mode photon synapses are connected, and single-mode waveguides of the multi-mode photon synapses are connected with single-mode waveguides of the multi-mode photon synapses; multimode waveguides of the second to Nth multimode photon synapses arranged along the column are connected; and N multi-mode photonic neurons connected with the last multi-mode waveguide of the corresponding column. According to the invention, a brand new mode division multiplexing integrated photon neural mimicry computing architecture is provided, the limitation of space stacking of a traditional single-mode device is broken through, and the optical parallel computing and neuron activation capability is really expanded by using a mode division multiplexing technology.

The embodiment of the specification discloses a large language model illusion path accurate blocking method and system. Among them, the method comprises: determining the model fact illusion output and the corresponding neuron activation sequence by performing fact illusion judgment on the model output of the large language model; determining the time-varying graph according to the neuron activation sequence; determining the illusion generation time window according to the model fact illusion output, and determining the node weighted degree of the time-varying graph in the illusion generation time window, the node weighted degree is used to determine the structural entropy contribution degree and the Shannon entropy; based on the gradient direction of the neuron activation based on the Shannon entropy, the illusion output word node is traced back; based on the candidate illusion path, the target illusion path is identified in the model reasoning process, and hierarchical accurate intervention is implemented through topological isolation technology. The embodiment of the specification can accurately block the model illusion path, and improve the model security and reliability.

The invention relates to the technical field of microscopic optical imaging, in particular to a pulse time-sharing multi-channel three-dimensional scanning imaging system, which comprises a laser pulse time-sharing module for equally dividing femtosecond laser into four paths and realizing time-sharing excitation through optical path difference adjustment; the four-way parallel tracking scanning module is used for scanning the width of the dendrite and tracking the trajectory, and imaging the curved dendrite into a straight two-dimensional strip; the axial multi-point imaging module is used for carrying out wavefront modulation by utilizing a spatial light modulator and axially generating a plurality of separated focus scanning depths on the focal plane of an objective lens; according to the invention, the information flux is improved through multi-component parallel scanning, high-speed high-resolution structure and functional imaging of multiple complete dendrites of the same neuron are realized, and the imaging precision is improved. Neuron multi-site input and output synchronous observation in a cognitive state is supported, and high-precision experimental data is provided for a neuron activation kinetic model.

The embodiment of the specification discloses a large language model illusion path accurate blocking method and system. Among them, the method comprises: determining the model fact illusion output and the corresponding neuron activation sequence by performing fact illusion judgment on the model output of the large language model; determining the time-varying graph according to the neuron activation sequence; determining the illusion generation time window according to the model fact illusion output, and determining the node weighted degree of the time-varying graph in the illusion generation time window, the node weighted degree is used to determine the structural entropy contribution degree and the Shannon entropy; based on the gradient direction of the neuron activation based on the Shannon entropy, the illusion output word node is traced back; based on the candidate illusion path, the target illusion path is identified in the model reasoning process, and hierarchical accurate intervention is implemented through topological isolation technology. The embodiment of the specification can accurately block the model illusion path, and improve the model security and reliability.

The application discloses a kind of based on decoupling neuron activation mechanism's power transmission and distribution related equipment protection method, equipment and medium.The method comprises the following steps: obtaining the image of power transmission and distribution related equipment and the request behavior data for power transmission and distribution related equipment;The image of power transmission and distribution related equipment is input into target trunk model, and the bottleneck layer feature vector output by target bottleneck layer in target trunk network and the output vector of target trunk network are obtained;According to bottleneck layer feature vector and request behavior data, determine target request vector;Determine the similarity between each abnormal request vector in abnormal request vector set and target request vector;Bottleneck layer feature vector is input into pre-trained activation model, and target confidence is obtained;According to the similarity between each abnormal request vector in abnormal request vector set and target request vector, determine target activation value;According to target activation value, target confidence and the output vector of target trunk network, determine the running state of power transmission and distribution related equipment.

The invention provides a diffusion model privacy data protection method based on resampling and anomaly detection. The method comprises the following steps: performing fine adjustment on a pre-trained text graph diffusion model; inputting a text prompt required to be forgotten by a user into the fine-tuned text graph diffusion model, and calculating a distribution condition of neurons concerned by the user in a corresponding neuron module; screening neurons of which the activation values are greater than a preset threshold value in the distribution; carrying out re-sampling operation on the screened neuron activation values according to original parameter distribution; according to the method, the screened neuron parameters are re-sampled by adopting a self-service method, and the original parameters are replaced by the parameter values obtained by sampling, so that efficient privacy data removal is realized on the premise of not damaging the generation quality and diversity of the model; the forgetting depth and speed of the model are optimized, and the privacy of the user is effectively protected; and the efficiency of the model is greatly improved, and the required calculation consumption is reduced.

The application provides a large model knowledge base construction method and system for traffic logistics, based on the standard term set of the traffic logistics field, semantic relationship mining is performed to generate a structured term atlas, a template knowledge probe set is constructed, a pre-trained large language model is input to perform directional activation detection operation, and an implicit knowledge neuron cluster is obtained, a spatiotemporal correlation analysis is performed on the activation response mode thereof, the distribution characteristics of the neuron activation intensity and the dependency relationship between the characteristics are extracted, the implicit knowledge neuron cluster is mapped to a semantic traffic logistics field knowledge unit, hierarchical organization and semantic linking are performed according to the logical correlation degree and the functional attribute classification result, and a structured knowledge network in the traffic logistics field is generated; a standard traffic logistics business problem set is called to verify the knowledge utility, the knowledge unit and the associated relationship are dynamically optimized and adjusted, a large model knowledge base is obtained, and high-quality knowledge support is provided for intelligent application in the traffic logistics field.

There are provided systems and methods for data privacy protection and removal for artificial intelligence model training and deployment. An online transaction processor or other service provider may provide computing services and platforms to entities, which may include use of machine learning (ML) models including large language models (LLMs). To comply with data privacy protections and copyright enforcement, a system may provide unlearning of content from ML models. The system may receive a request to unlearn a content and, after verifying the request is valid, identify the content used for during training of or inferencing by an ML model. The system may then map the content to concepts and correlate those concepts with ML model outputs using projections in a vector space. Based on the mapped concepts and outputs, neuron activation of the ML model may be analyzed to identify a negation vector and perform selective parameter dampening.

The present invention is directed to antibodies and antigen binding fragments thereof having binding specificity for PACAP. The antibodies and antigen binding fragments thereof comprise the sequences of the VH, VL, and CDR polypeptides described herein, and the polynucleotides encoding them. Antibodies and antigen binding fragments described herein bind to and / or compete for binding to the same linear or conformational epitope(s) on human PACAP as an anti-PACAP antibody. The invention contemplates conjugates of anti-PACAP antibodies and binding fragments thereof conjugated to one or more functional or detectable moieties. Methods of making said anti-PACAP antibodies and antigen binding fragments thereof are also contemplated. Other embodiments of the invention contemplate using anti-PACAP antibodies, and binding fragments thereof, for the diagnosis, assessment, and treatment of diseases and disorders associated with PACAP and conditions where antagonism of PACAP-related activities, such as vasodilation, photophobia, mast cell degranulation, and / or neuronal activation, would be therapeutically beneficial.

The application relates to the technical field of deep learning and artificial intelligence model optimization, and discloses a model merging method and device, an electronic device and a storage medium, the method comprising the following steps: constructing a preset dimension task tensor based on task vectors of a plurality of preset expert models and a pre-training base model; performing Tucker decomposition on the task tensor to obtain a core tensor and a factor matrix; constructing a static correction matrix based on a residual tensor of the Tucker decomposition and a task mask; the task mask is generated by analyzing neuron activation patterns in each preset expert model; in a model inference stage, a dynamic update matrix is reconstructed based on real-time input data, the factor matrix and the core tensor, and the dynamic update matrix and the static correction matrix are merged with the pre-training base model to obtain a fused model parameter, the application realizes efficient merging of multiple expert models without accessing original training data, reduces deployment cost, and improves multi-task processing performance.

Systems and methods for detecting a physiological response to neuronal activation include a wearable device configured to be worn on a portion of an arm of a user. The wearable device includes a plurality of electrodes configured to detect biopotential signals from the user's arm and a processor coupled to the plurality of electrodes. The processor is configured to analyze biopotential data derived from the biopotential signals to determine an activation level of the user and generate an output indicating the activation level of the user. The system may include additional sensors such as an inertial measurement unit, a photoplethysmography sensor, a galvanic skin response sensor, a temperature sensor, an electrocardiogram sensor, and an ambient light sensor to provide complementary data for analysis. The processor may classify the activation level, trigger actions based on the determined level, and correlate the activation level with cognitive performance.

The embodiment of the specification discloses a large model output result detection method. The method is based on training samples labeled with whether the output result is correct, and determines the activation state of neurons in the inference process of the large model LLM to be detected, to obtain the relationship between the correctness of the output result and the activation state of the neurons, and train a detection classifier. Then, when the LLM to be detected executes task input data, the activation state of the neurons is obtained, and the correctness of the LLM output result is detected by the detection classifier, that is, whether hallucination occurs. The method links the relationship between the activation state of the LLM neurons and the occurrence of hallucination of the LLM, so as to identify whether the LLM hallucinates, without paying attention to the inference process and the output result of the LLM, and identifying whether the LLM hallucinates from the bottom of the model.

This invention relates to the field of edge computing and federated learning security technology, specifically disclosing a lightweight secure aggregation and adversarial backdoor defense method based on an edge AI model. The server receives sparse updates from edge nodes and performs weighted aggregation after hash verification to obtain a global model. The global model is then forward-propagated on clean samples, dynamically expanding the samples until the variance of the activation value sequence converges, obtaining the activation value distribution of each neuron. The skewness, kurtosis, and median absolute deviation of the distribution are calculated. Neurons with constant activation values ​​are identified as backdoor neurons and pruned, while high-skewness, high-peak false positive neurons with domain offset are retained. After pruning, the robust average activation value of the retained neurons is calculated as a baseline vector and sent to the edge nodes. The edge nodes calculate the deviation between their local actual activation and the baseline and sum them as a bias term before starting local training. This invention can distinguish between domain offset and real backdoors, avoid false pruning, and improve the security of edge federated learning.