382 results about "Neural network system" patented technology

A computer-implemented method of training a neural network system comprising a visual encoder neural network and a text encoder neural network is provided The method comprises obtaining a plurality of training data items (each training data item comprising an image and associated text defining a sequence of text tokens) and at each of a plurality of training steps processing at least one of the training data items by: processing pixels of the image in the training data item using the visual encoder neural network to generate a set of patch embeddings for the image; processing the sequence of text tokens using the text encoder neural network to generate a sequence of token embeddings, processing the set of patch embeddings and the sequence of token embeddings to generate a set of language-aware patch embeddings (based on similarities between patch embeddings and token embeddings), and training at least the visual encoder neural network by backpropagating gradients of a contrastive objective function evaluated over the language-aware patch embeddings and the sequence of token embeddings.

A neural network system is provided. The system includes an autoencoder configured to encode input data into a latent space representation; a generator neural network configured to receive a noise vector and the latent space representation and output a set of routing coefficients; a discriminator neural network configured to evaluate the effectiveness of the routing coefficients by measuring the performance of a capsule network utilizing said routing coefficients; and a capsule network comprising a first capsule layer and a second capsule layer, wherein the routing coefficients are used to dynamically route outputs from the first capsule layer to the second capsule layer.

A method for robot path planning using skills extracted from human-taught motion programs applied to a new obstacle environment. A three-dimensional convolutional neural network is used to extract features characterizing an obstacle environment, where the feature vector representation of the obstacles overcomes problems encountered when using point cloud obstacle data. The obstacle feature data and robot path start and goal points are provided to an encoder / decoder neural network system which is trained to extract skills from a database of human-generated motion programs. The encoder / decoder neural network system produces a distribution of waypoints for the current obstacle environment and start / goal points. The distribution of waypoints is used to perform a final collision-free path generation using either a rapidly-exploring random tree (RRT) technique or an optimization-based technique.

A neural network system is trained to generate textual reports (that is, medical reports, such as radiology reports) from one or more medical images, by fine-tuning a pre-trained neural network system (a visual language model, "VLM") operative, upon receiving an input comprising at least one image and a textual input, to generate a value indicative of a predicted likelihood of one or more candidate text continuations of the textual input. The fine-tuning of the neural network system is performed to reduce the value of a cost function which includes a first prediction cost term based on a first training database including first training datasets of at least one medical image and an associated text report, the first training datasets corresponding to first individuals. The first prediction cost term further includes a cost value for each individual, inversely dependent on a likelihood value of the associated textual report, conditioned on the at least one medical image, and created by the neural network system.

A neural network system that is configured to learn a representation of data item, such as an image, audio, or text data item, through a self-supervised learning process. Implementations of the system couple two learning processes, an object discovery learning process and an object feature representation learning process. In implementations the object discovery learning process assists the object feature representation learning process in self-supervised learning of object feature representations, and the object feature representation learning process is used to improve the object discovery learning process.

The invention discloses an intelligent tea identification and grading method and system based on a neural network, and relates to the technical field of intelligent tea identification and grading, in-depth learning and feature extraction are performed on tea images through a neural network classification module and a convolutional neural network, and fine features such as color, shape and size of tea can be identified. And high-precision grading is carried out. Compared with traditional artificial grading, the neural network system eliminates interference of artificial deviation and subjective factors, and ensures consistency and reliability of grading results. Through adaptive filtering and image enhancement algorithms in the image preprocessing module, the system can automatically carry out illumination correction, noise removal, background removal, image sharpening and other processing in the image acquisition stage, and the definition and quality of the image are greatly improved. According to the system, through automatic image acquisition, processing, feature extraction and grading processes, manual intervention and operation can be greatly reduced, and the tea grading efficiency and speed are improved.

A method is provided for processing data in a neural network system. The method includes receiving input data; processing the input data through a first set of neural network layers configured to perform data processing using MatMul-free techniques to produce intermediate data; further processing the intermediate data through a second set of neural network layers configured to simulate spiking neural network (SNN) functionalities using MatMul-free techniques; and outputting a result based on the processed data from the second set of neural network layers.

The invention discloses a virtual power plant optimal scheduling method based on artificial intelligence, and further reduces the power operation cost through an integrated intelligent management and scheduling strategy. The system is composed of a power generation end, an energy storage end, a power utilization end, a system management platform, a power grid, a power market and other key components. The core lies in that a depth deterministic strategy gradient (DDPG) algorithm is utilized to construct a simulation environment model, and a reward function is designed to evaluate the economic benefit of the strategy. By initializing Actor and Critic neural networks, the system can generate and evaluate a scheduling strategy, and continuously iteratively updates the network through an experience playback mechanism to optimize economic benefits. The invention provides an innovative virtual power plant optimal scheduling method based on artificial intelligence, and maximization of the benefits of the virtual power plant is realized by reducing the operation cost.

Systems and methods for performing image modification. In particular, the system can, using a rectified flow neural network, perform an image inversion and image editing process to generate a modified image that has been modified according to a conditioning input received by the system.

Systems and methods are provided for analog hardware realization of neural networks. The method includes obtaining a neural network topology and weights of a trained neural network. The method also includes transforming the neural network topology to an equivalent analog network of analog components including operational amplifiers and resistors. Each operational amplifier represents an analog neuron of the equivalent analog network, and each resistor represents a connection between two analog neurons. The method also includes computing a weight matrix based on the weights of the trained neural network. The method also includes generating a resistance matrix for the weight matrix. The method also includes pruning the equivalent analog network to reduce the number of operational amplifiers or the resistors, based on the resistance matrix, to obtain an optimized analog network of analog components.

According to various embodiments, machine learning techniques for determining a pharmacokinetic parameter of a radiotracer in a patient are presented. The techniques include: obtaining a time sequence of positron emission tomography (PET) scans of a target tissue of the patient, where the time sequence of PET scans includes images of a radiotracer in the target tissue; providing the time series of the PET scans to a first neural network system, which provides features extracted from the time series of the PET scans; providing the features extracted from the time series of the PET scans to a second neural network system, which provides an estimated arterial input function; modeling the pharmacokinetic parameter of the radiotracer based on the estimated arterial input function and the time sequence of PET scans; and outputting an estimate of the pharmacokinetic parameter of the radiotracer based on the modeling.

The invention relates to the technical field of geological early warning, in particular to a dangerous rock mass instability analysis method, system and equipment based on a space-time diagram neural network, by fusing unmanned aerial vehicle LiDAR, multispectral data, meteorological radar data and the space-time diagram neural network (ST-GNN), the system realizes sub-meter spatial resolution and minute-level time response, and the stability of dangerous rock mass instability analysis is improved. The four-dimensional (time and space) analysis result of the instability probability of the dangerous rock mass is obtained through high-precision space-time modeling, the problems that a traditional geological disaster early warning system is low in resolution ratio, slow in response and high in misinformation are solved, the comprehensiveness, accuracy and reliability of instability prediction of the dangerous rock mass are improved, and the early warning effect is good. And full-chain intelligent closed-loop management of real-time data acquisition-dynamic prediction-early warning push-feedback optimization is supported, the emergency decision time is shortened by real-time rainfall superposition risk thermodynamic diagrams, and the attenuation rate of long-term prediction precision is reduced by dynamically fusing newly added geological data and instability events through incremental learning.

The invention relates to an event trigger pulse control method and system of a neural network under a fraud attack, and belongs to the technical field of neural network control. Comprising the following steps: establishing a coupling neural network model containing time variation and time delay, and determining a synchronization target of the coupling neural network model; acquiring state information of the coupled neural network model, and establishing an error coupled neural network system model according to the state information; designing a saturation pulse controller according to the error coupling neural network system model and the synchronization target; obtaining a first sufficient condition and a second sufficient condition based on the error coupling neural network system model and the saturation pulse controller; obtaining a first optimization mathematical model according to the first sufficient condition and the second sufficient condition; and calculating by using the first optimization mathematical model to obtain the maximum estimation of the pulse control gain matrix and the attraction domain. According to the method, the joint influence of actuator saturation, spoofing attack and pulse effect on the overall performance of the system is effectively solved.

Systems and techniques are described herein for training and using a machine-learning model (e.g., a neural network). For example, a computing device can: process, using a first trained neural network, data specific to a user to obtain intermediate activation data representing the data, the first trained neural network comprising a plurality of neural network layers; process, using a second trained neural network, the intermediate activation data to generate an output representing the data, the second trained neural network comprising a subset of neural network layers from the plurality of neural network layers of the first trained neural network; determine a loss based on the output; and update parameters of the second trained neural network based on the loss.

The invention relates to the technical field of remote maintenance of a power distribution cabinet, in particular to a remote maintenance system of an intelligent power distribution cabinet, which constructs a component operation parameter matrix by collecting current, voltage, component temperature, starting time and internal and external environment data of the power distribution cabinet in real time, and combines environment parameters of the power distribution cabinet to maintain the power distribution cabinet. And carrying out deep learning training by using a residual neural network algorithm. The method not only can detect common fault modes, but also can identify abnormal conditions caused by component aging, so that the problems of false alarm and missing alarm in a traditional fault detection method are solved. Through the residual neural network, the system can automatically learn normal and abnormal operation modes from historical data, and the accuracy of fault detection is improved. And the exception type is further pushed to the operation and maintenance personnel and is visually displayed, so that the judgment time and the round-trip times of the operation and maintenance personnel on the maintenance component are saved, and the maintenance efficiency is improved.

The invention belongs to the field of hardware neural network design, and particularly discloses a CMOS-based attractor recurrent neural network system and an implementation method thereof.A neuron calculation module receives an external input high-level signal and generates an initial synaptic high-level signal; integration and activation calculation are carried out based on the current output by the synaptic calculation module, and whether a synaptic high-level signal is generated or not is determined according to an activation state; the synaptic storage module writes the non-zero weight into a synaptic weight SRAM (Static Random Access Memory) array; the synaptic calculation module performs product operation on the synaptic high-level signal and the synaptic weight in the synaptic weight SRAM array, and outputs current; when the STDP updating module is in a neuron training mode, the STDP updating module updates the synaptic weight SRAM array according to the synaptic high-level signal; and the group state module outputs a neuron sequence number corresponding to the maximum activation frequency. A cyclic connection structure of the attractor neural network is formed, and the real-time cognitive behavior requirement of the brain is met.

Systems and methods for using a distributed computing system to train a large neural network to perform a machine learning task. A shared set of trainable parameters is maintained in a shared data store, and each of a geographically distributed set of workers updates their trainable parameters using a shard training dataset. There are two optimization processes: an outer optimization process, and an inner optimization loop that is executed by each worker independently and in parallel tens, hundreds, or thousands of times. The workers can have different computing capabilities and can be geographically distant from one another, and the communications bandwidth used by the system can be two or three orders of magnitude less than that of other systems.

Techniques are described herein for processing video data using a neural network system. For instance, a process can include generating, by a first convolutional layer of an encoder sub-network of the neural network system, output values associated with a luminance channel of a frame. The process can include generating, by a second convolutional layer of the encoder sub-network, output values associated with at least one chrominance channel of the frame. The process can include generating, by a third convolutional layer based on the output values associated with the luminance channel of the frame and the output values associated with the at least one chrominance channel of the frame, a combined representation of the frame. The process can further include generating encoded video data based on the combined representation of the frame.

A computer-implemented method for training an anomaly detection neural network system comprising an encoder and a decoder is described. The method includes receiving training data comprising a plurality of training examples, each training example including a training audio waveform and a machine identity (ID); processing the training audio waveform to extract training audio features; receiving an environmental noise audio waveform; processing the environmental noise audio waveform to extract noise features; generating augmented features by combining the extracted training audio features and the noise features; processing, using the encoder, the augmented features to generate latent embeddings; processing, using the decoder, the latent embeddings to generate reconstructed audio features; processing, using a convolutional neural network, the augmented features to generate a predicted machine ID probability distribution; and adjusting, through backpropagation, the current values of the parameters of the encoder and the decoder to minimize an objective function.

The invention discloses an ultra-lightweight speech enhancement neural network system and method based on adaptive convolution. The system comprises an encoder; a frequency spectrum compression module, which is used for carrying out dynamic range and frequency dimension compression on a noisy speech frequency spectrum; the self-adaptive convolution attention module is used for carrying out frequency dimension downsampling and feature extraction analysis on the compressed feature map through a self-adaptive convolution layer and time channel attention, and then sending the feature map into an intensifier; the intensifier is a dual-path packet recurrent neural network and is used for modeling sub-band time features and intra-frame frequency features of the voice; the decoder comprises a self-adaptive transpose convolution attention module which is used for performing frequency dimension upsampling and feature reconstruction on the feature map output by the enhancer; and the frequency spectrum expansion module is used for expanding the frequency dimension of the feature map output by the self-adaptive transpose convolution attention module. According to the invention, high-performance speech enhancement can be realized with extremely low calculation complexity.

A method for robot path planning uses skills extracted from a motion program of human teaches applied to a new obstacle environment. A three-dimensional convolutional neural network is used to extract features representing an obstacle environment, wherein feature vector representation of an obstacle overcomes problems encountered when point cloud obstacle data is used. Obstacle feature data and robot path start and target points are provided to an encoder / decoder neural network system trained to extract skills from a database of human-generated motion programs. An encoder / decoder neural network system generates a waypoint distribution for a current obstacle environment and a start / target point. The waypoint distribution is used to perform final collision-free path generation using a fast exploring random tree (RRT) technique or an optimization-based technique.

A computer-implemented method for generating and ranking source code for performing a task is described. The method includes: receiving input data comprising a task description, a code generation prompt and a test case generation prompt; processing the input data using at least one trained code generation neural network to generate a plurality of code solutions and a plurality of test cases; for each code solution, executing the set of candidate source code on the test inputs of the plurality of test cases to generate a plurality of execution outputs; clustering the plurality of code solutions into a plurality of clusters; computing an interaction matrix that specifies functional overlap between the plurality of clusters; determining, for each cluster, a score based the interaction matrix; and ranking the plurality of clusters based on the scores of the plurality of clusters.

The invention provides an under-actuated unmanned ship trajectory tracking control method, which is used for under-actuated unmanned ship trajectory tracking control with input quantization and output limitation based on an event trigger mechanism, and comprises the following steps: acquiring sea condition information of a surrounding environment and other ships, and establishing a trajectory tracking control model of an unmanned ship; a composite quantizer is adopted to quantize control input of a control system, and a linear analysis model is used to describe an input quantization process; designing a barrier Lyapunov function on the basis of an output limited theory, compensating uncertain items and external interference in the under-actuated unmanned ship model by using a neural network system, and obtaining a trajectory tracking controller and an adaptive rate of the under-actuated unmanned ship; an event triggering mechanism is introduced, and the control rate of the under-actuated unmanned ship system is updated based on an event triggering strategy; it is proved that the stability of the designed under-actuated unmanned ship trajectory tracking control system is finally bounded when priori information of parameters does not need to be quantized, and all signals are finally bounded, so that Zeno behaviors are avoided.

Methods, systems, and apparatus, including computer programs encoded on computer storage media, for processing inputs using a neural network system that includes one or more pre-normalized layers or one or more regularization normalization layers.

Systems and methods, implemented as computer programs on one or more computers in one or more locations, for generating a sequence of data elements using a neural network comprising a sequence of attention neural network layers. The sequence comprises a respective continuous valued data element at each position in a sequence of positions. Implementations of the described techniques remove the need for discrete tokens and fixed, finite vocabularies.

The present invention discloses a memristor pulse neural network system and method, which belongs to the field of artificial intelligence technology. The memristor pulse neural network system provided by the present invention uses a bionic design concept to accurately simulate the synaptic structure of a biological neural network by connecting pairs of memristors of the same level in series; based on the positive and negative Hebbian plasticity rules, an adaptive learning mechanism is implemented to strengthen the computing function of the memristor's storage and computing performance; a cascaded scalable control architecture combining master control and slave control is adopted to give the system a high degree of scalability. According to the needs of specific tasks, users can flexibly adjust the scale of the neural network to achieve unlimited cascade expansion, thereby easily coping with various complex scenarios. In addition, the built-in ADC acquisition module provides a real-time feedback mechanism, which can improve the accuracy of system monitoring, speed up the response speed, and provide valuable data support for system optimization and adjustment.

One aspect of the technology includes recovering image affine transforms via CNN-based classification and regression. Another aspect of the technology is a CNN-based network to detect a presence / absence of a digital watermark signal and recovery of affine transform coefficients associated with an image template. Other aspects, features and arrangements are also described and claimed.

Systems and techniques are provided for multi-stage training of a multi-network system. An example method can include training, using training data, a first neural network during a first training stage; generating, by the first neural network, one or more outputs; based on a determination that the first training stage and training of the first neural network are complete, providing the one or more outputs to a second neural network that has an input data dependency comprising data generated by the first neural network; and based on the determination that the first training stage and training of the first neural network are complete, training, using the one or more outputs from the first neural network, the second neural network during a second training stage initiated after completion of the first training stage.

Systems and method are provided that are directed to tuning a hyperparameter associated with a small neural network model and transferring the hyperparameter to a large neural network model. At least one neural network model may be received along with a request for one or more tuned hyperparameters. Prior to scaling the large neural network, the large neural network is parameterized in accordance with a parameterizing scheme. The large neural network is then scaled and reduced in size such that a hyperparameter tuning process may be performed. A tuned hyperparameter may then be provided to a requestor such that the hyperparameter can be directly input into the large neural network. By tuning a hyper parameter using a small neural network, significant computation cycles and energy may be saved.

The application discloses a neural network system, a method, and a device for processing floating-point number. A self-defined floating-point number is obtained, wherein the self-defined floating-point number comprises a sign field, an exponent field, and a mantissa field, and a value of the self-defined floating-point number is determined by bit of the sign field, bits of the exponent field, bits of the mantissa field, and a bias value, wherein the bias value is determined by a total bit number of the exponent field. The self-defined floating-point number is applied to numerical calculations.