102 results about "Artificial neural network" patented technology

Disclosed herein is an artificial neural network-based medical image diagnosis assistance apparatus for assisting in diagnosing a medical image based on a medical artificial neural network. A medical image diagnosis assistance apparatus according to an embodiment of the present invention includes a computing system, and the computing system includes at least one processor. The at least one processor is configured to acquire or receive a first analysis result obtained through the inference of a first artificial neural network about a first medical image, to detect user feedback on the first analysis result input by a user, to determine whether the first analysis result and the user feedback satisfy training conditions, and to transfer the first analysis result and the user feedback satisfying the training conditions to a style learning model so that the style learning model is trained on the first analysis result and the user feedback.

Disclosed is a computational fluid dynamics (CFD) flow analysis method for blades includes: generating an analysis automation model that predicts CFD flow analysis results according to input analysis conditions, the generating an analysis automation model comprising inputting a blade shape and flow conditions, determining analysis settings for flow analysis over blades, performing CFD simulation based on the determined analysis setting parameters to conduct flow analysis over blades, evaluating results of the performed flow analysis over blades, and training through reinforcement learning using artificial neural networks to satisfy predetermined evaluation criteria, inputting arbitrary blade shape conditions and flow conditions, and determining analysis setting parameters by applying the analysis automation model to the input blade shape and flow conditions, predicting CFD simulation, and outputting the predicted CFD simulation as flow analysis results.

A processor-implemented method performed for text-based video editing includes receiving a video input and a text prompt. The video input includes a sequence of video frames. Features of the video input are extracted to generate a latent representation of the video input. Noise is injected to the latent representation of the video input to generate a noise injected latent. The noise is conditioned on the video input. An artificial neural network (ANN) model processes the noise injected latent based on the text prompt to adapt the video input according to the text prompt.

An artificial neural network (ANN) receives a sequence of input data tokens. The ANN has a hidden state memory of size N and a set of weights having a size that scales on or below an order of N. The ANN processes each token in the sequence of input data tokens by performing, based on each token, a logical operation on the hidden state memory to generate an updated hidden state memory. Processing the sequence of input data tokens comprises performing a number of compute operations that scales on or below an order of N{circumflex over ( )}1.5. The ANN obtains, from the updated hidden state memory after processing each token, a final hidden state memory. The ANN generates an inference result based on the final hidden state memory.

Systems and methods for developing a game of chance. The game of chance being at least partially developed using specialized artificial intelligence (AI) game design systems or specialized artificial intelligence game design system modules or components which may include different types of machine learning or training techniques, including supervised, unsupervised, reinforced, deep learning and artificial neural networks and / or similar and may include analyzing past game performance utilizing full, partial or estimated prior game performance data for developing slot machine game math models, slot machine game mechanics and associated game programming, slot machine game art and graphics, slot machine animations, slot machine game sound effects, partial or full slot machine game development, slot machine computer code, slot machine game quality assurance diagnosis and editing, slot machine game analytics, slot machine compliance, slot machine help screens, and slot game predictive models, etc.

The disclosed computational fluid dynamics (CFD) airfoil flow analysis method comprises: an analysis automation model generation step for inputting the shape and flow conditions of a blade, determining analysis settings for airfoil flow analysis, performing the airfoil flow analysis by performing a computational fluid dynamics simulation on the basis of parameters for the determined analysis settings, and generating a model for predicting a CFD flow analysis result according to input analysis conditions by evaluating the results of the performed airfoil flow analysis to train the model through reinforcement training using an artificial neural network so as to satisfy a preset evaluation criterion; an analysis condition input step for inputting a shape condition and flow condition of an arbitrary blade; and an analysis result output step for determining the analysis setting parameters by applying the analysis automation model to the input shape condition and flow condition of the blade and performing CFD simulation to output the predicted CFD flow analysis result.

A processor-implemented method includes obtaining a color filter array (CFA) input image, obtaining pattern information corresponding to a CFA, preprocessing the input image based on the pattern information, generating an inferred image by inputting the preprocessed input image to an artificial neural network (ANN) model, and generating an output image by selecting, for each pixel, from either one of the preprocessed input image and the inferred image, based on the pattern information.

A method for detecting defects for an aeronautical part, includes training, during a plurality of epochs, an artificial neural network to supply a defect probability for each pixel of an image, for each epoch and for each image of a validation set, creating a defect probability matrix based on defect probabilities for each pixel of the image, for each epoch, determining, for each image of the validation set, based on the defect probability matrices, a curve of defect detection as a function of false alarms, determining a set of the N best epochs on the basis of the defect detection curves and of a business criterion, for each epoch of the N best epochs, inspecting the image to be inspected by applying the artificial neural network with parameters associated with the epoch, so as to obtain an inspected sub-image associated with the training epoch.

Computer-implemented machines, systems and methods for providing insights about misalignment in a latent space of a machine learning model. A method includes initializing a second weight matrix of a second artificial neural network based on a first weight matrix from a first artificial neural network. The method further includes applying transfer learning between the first artificial neural network and the second artificial neural network. The method further includes comparing the first latent space with the second latent space. The method further includes determining, responsive to the comparing, a first score indicating alignment of the first latent space and the second latent space. The method further includes determining, and responsive to the first score satisfying a threshold, an appropriateness of the machine learning model.

A method for processing image data is provided. The method including: providing source data from images acquired by a camera viewing an object, providing target data related to the object, operating an AI system including at least one artificial neural network, wherein the source data are used as an input of the AI system, and wherein the AI system is trained to identify a first correlation between the source data and the target data, and processing the source data and the target data based on the first correlation to obtain a second correlation between the source data and the target data.

A data processing apparatus comprises a captioning model to receive gameplay telemetry data indicative of one or more in-game properties for a session of a video game, the captioning model comprising an artificial neural network (ANN) trained to output caption data comprising one or more captions in dependence upon a learned mapping between gameplay telemetry data and caption data, one or more of the captions comprising one or more words for providing a visual description for the session of the video game, and output circuitry to output one or more of the captions.

A machine learning system is provided to enhance various aspects of machine learning models. In some aspects, a substantially photorealistic three-dimensional (3D) graphical model of an object is accessed and a set of training images of the 3D graphical mode are generated, the set of training images generated to add imperfections and degrade photorealistic quality of the training images. The set of training images are provided as training data to train an artificial neural network.

A processor-implemented method for estimating uncertainty includes receiving an artificial neural network (ANN) model and a calibration dataset. An upper-bound on a conditional entropy of a data distribution of the ANN model is computed using conformal prediction. The conditional entropy of the model is estimated based on the upper-bound. In turn, the uncertainty of the data distribution of the ANN model may be estimated based on the conditional entropy and a prediction set size.

A method of wireless communication, by a user equipment (UE) includes receiving, from a base station, a jointly trained artificial neural network. The method also includes calculating a value representing at least one of (1) a gradient estimate for a weight of the jointly trained artificial neural network, or (2) the weight of the jointly trained artificial neural network. The method further includes expanding the value into a numerical system with base N into a plurality of digits. The method also includes determining a number and / or a location of the plurality of digits to transmit based on a deterministic task assignment rule received from the base station or a probabilistic task assignment rule. The method further includes transmitting the determined number and / or the determined location of the plurality of digits to the base station.

A processor-implemented method for video compression using an artificial neural network (ANN) includes receiving a video via the ANN. The ANN extracts a first set of features of a current frame of the video and a second set of features of a reference frame of the video. The ANN determines an estimate of correlation features between the first set of features of the current frame and the second set of features of the reference frame. The estimate of the correlation features are encoded and transmitted to a receiver.

A system and method are disclosed for estimating the floor area of an existing building from a 3D digital representation. Various embodiments can leverage reality capture devices, such as LIDAR laser scanners or photogrammetry techniques, to obtain 3D representations. The 3D representations may be segmented using artificial neural networks, isolating individual floors. These segmented floors may be projected into 2D and processed through another neural network to yield a precise binary representation differentiating between floor and non-floor areas. The resulting binary images allow for the computation of the total floor area in square footage for each level.

Methods, apparatus and systems that relate to hardware accelerators of artificial neural network (ANN) performance that significantly reduce the energy and area costs associated with performing vector dot-product operations in the ANN training and inference tasks. Specifically, the methods, apparatus and systems reduce the cost of bit-level flexibility stemming from aggregation logic by amortizing related costs across vector elements and reducing complexity of the cooperating narrower bitwidth units.

The present invention relates to a method and a device of inputting annotation of object boundary information, and more particularly, to a method and a device of inputting annotation of object boundary information such that workers or the like efficiently and accurately input object boundary information in a service such as crowding sourcing, and preliminary work is assisted by an artificial neural network-based algorithm.

A data compression system (119) configured to perform data compression of input data to generate compressed input data. Performing the data compression comprises: receiving (205) input data; dividing (207) the input data into a plurality of input chunks; operating a large language model to generate (215), for each of the plurality of input chunks, a reduced summary of the input chunk; and collating (217) the plurality of reduced summaries to generate the compressed input data. Also disclosed is a generative artificial intelligence system comprising an artificial neural network (137) configured to receive the compressed input data and one or more questions, and generate, on the basis of the compressed input data, a response to each of the one or more questions. The division of the input data into a plurality of input chunks may be done according to one or more contextual boundaries of the data. Examples of contextual boundaries include paragraph, page, or section breaks, or different topics. Contextual boundaries may be identified by vectorising the data. The reduced summary may be produced by selectively discarding data from the respective input chunk.

An apparatus and method for diagnosing developmental disorder using an artificial neural network are disclosed. An apparatus for diagnosing developmental disorders according to one exemplary embodiment includes a communication unit that performs communication with one or more user terminals and a control unit connected to the communication unit, in which the control unit receives at least one of image data and voice data for an analysis target from the user terminal and performs a first developmental disorder diagnosis through one or more artificial neural networks learned to perform a developmental disorder diagnosis, receives questionnaire response data from the user terminal and performs a second developmental disorder diagnosis according to predetermined standard, and generates developmental disorder diagnosis data for the analysis target based on the first developmental disorder diagnosis and the second developmental disorder diagnosis.

An intelligent (self-learning) sensor-aware and / or context-aware subsystem comprising (i) a System-on-a-Chip (SoC), (ii) a radio transceiver, (iii) a microphone, (iv) a voice processing module, (v) a (bio-inspired) neuromorphic event camera or a hyperspectral camera, (vi) a digital personal assistant (DPA), (vii) a first set of computer implemental instructions in artificial neural networks (ANN) (which may include a transformer model / diffusion model or Poisson flow generative model++ (PFGM++)), (viii) a second set of computer implementable instructions to analyze and interpret contextual data and (ix) an autonomous artificial intelligence (AI) agent is disclosed.

The methods and systems are directed to automated computer analysis and machine learning. Specifically, the systems and methods for using machine learning to generate fault prediction models and applying the fault prediction models to logical circuits to optimize test point insertion determinations and optimize fault detection in the logical circuit. Disclosed are methods and systems that that generates training data from training circuits (and optionally generate training circuits), trains a learning segment (which may include an artificial neural network (ANN)) using the training data. The learning segment (once trained) generates fault prediction models to predict the quality of a TP inserted on a given circuit location and optimize TPI for a given circuit. The methods and systems described provide computational (CPU / processing) time advantages and precision over conventional methods.

An artificial neural network memory system includes at least one processor configured to generate a data access request corresponding to an artificial neural network operation; and at least one artificial neural network memory controller configured to sequentially record the data access request to generate an artificial neural network data locality pattern of the artificial neural network operation and generate an advance data access request which predicts a next data access request of the data access request generated by the at least one processor based on the artificial neural network data locality pattern.

Aspects of the present invention are directed to computer-implemented techniques for improving the training of artificial neural networks using a reduced precision (e.g., float16) data format. Embodiments of the present invention rescale tensor values prior to performing matrix operations (such as matrix multiplication or matrix addition) to prevent overflow and underflow. To preserve accuracy throughout the performance of the matrix operations, the scale factors are defined using a novel data format to represent tensors, wherein a matrix is represented by the tuple X, where X=(a, v[.]), wherein a is a float scale factor and v[.] are scaled values stored in the float16 format. The value of any element X[i] according to this data format would be equal to a*v[i].

Testing circuitry and methods are disclosed for use with analog neural memory in deep learning artificial neural networks. The analog neural memory comprises one or more arrays of nonvolatile memory cells. The testing circuitry and methods can be utilized during sort tests, qualification tests, and other tests to verify programming operations of one or more cells.

Device with which a vehicle (101) is equipped, the device comprising: a receiver (155) of a navigation system; a camera (108); a memory (145); and a processor (140) coupled to the receiver (155) of the navigation system, the memory (145) and the camera (108), wherein the processor (140) is configured to execute a method for detecting one or more three-dimensional structures in an environment of the vehicle (101) at runtime, the method comprising: Generating, by the processor (140), using the camera (108), a bird's-eye view (BEV) camera image (204) of the environment of the vehicle (101), wherein the BEV camera image (204) includes two-dimensional coordinates of one or more structures in the environment; Generating, by the processor (140), a BEV height image of the environment of the vehicle (101), wherein the BEV height image provides a height of one or more structures in the environment; Detect, by the processor (140), one or more edges (402) of one or more three-dimensional structures in the environment based on the BEV camera image (204) and the BEV height image; Generating, by the processor (140), models of the one or more three-dimensional structures in the environment of the vehicle (101) by plane fitting based on the boundaries (402) of the one or more three-dimensional structures; and Reconfiguring, by the processor (140), of the receiver (155) of the navigation system based on the models of one or more three-dimensional structures in the environment of the vehicle (101); wherein the BEV height image is generated based on radar data acquired by a radar (110) of the vehicle (101); wherein the BEV elevation image is a BEV radar image (206) generated using radar data; including the detection of the edges (402) of one or more three-dimensional structures: Generating, by the processor (140), a merged image by combining the BEV camera image (204) and the BEV radar image (206) to add the elevation information as a channel for each pixel in the BEV camera image (204); and Detect, by the processor (140), the edges (402) by inputting the merged image into an artificial neural network.

Various techniques are provided for providing neural networks with increased efficiency. In one example, a system includes a first artificial neural network (ANN), a second ANN, and a logic device. The first ANN is configured to receive a first plurality of data inputs associated with a data stream and process the first data inputs to generate a first inference output after a first latency. The second ANN is configured to receive a second plurality of data inputs associated with the data stream and process the second data inputs to generate a second inference output after a second latency less than the first latency. The logic device is configured to receive the second inference output before the first inference output is generated. Additional systems and methods are also provided.

A method and a system for validating a trained artificial neural network on the basis of a test data set.

An artificial neural network for processing low-resolution images to generate super-resolution images includes feed-forward connections between layers. The network includes an input layer, one or more convolution layers, wherein the input layer is connected to a first convolution layer of the convolution layers, and a output layer connected to a last convolution layer of the convolution layers. An input image is input to the input layer and to at least one of the convolution layers, an initial output of the input layer is input to at least one of the convolution layers, and a layer output of at least one of the convolution layers is input to at least one subsequent convolution layer.

The invention discloses a direct-current micro-grid energy storage system performance balance droop control method, which comprises the following steps of: firstly, evaluating the residual capacity of each energy storage unit battery on line in real time based on an ANN battery capacity evaluation model; secondly, calculating a battery performance index based on the residual capacity of the battery, and calculating a performance index deviation degree of each energy storage unit battery; determining a system working mode according to the maximum deviation degree of the battery performance indexes of all the energy storage units, and starting droop control based on performance balance when the deviation of the battery performance indexes is large; next, each energy storage unit corrects the initial droop coefficient according to the battery performance index deviation degree to obtain an optimized droop coefficient, then a reference value of the battery output voltage of each energy storage unit is obtained through droop control, and finally a PWM switching signal is generated through voltage-current double closed-loop control of an energy storage unit interface DC / DC converter to enable the output to follow the reference value. Through the method, the overall service life of the direct-current micro-grid energy storage system can be effectively prolonged.