Abstract: A computer-implemented method includes receiving, by a computing device, input activations and determining, by a controller of the computing device, whether each of the input activations has either a zero value or a non-zero value. The method further includes storing, in a memory bank of the computing device, at least one of the input activations. Storing the at least one input activation includes generating an index comprising one or more memory address locations that have input activation values that are non-zero values. The method still further includes providing, by the controller and from the memory bank, at least one input activation onto a data bus that is accessible by one or more units of a computational array. The activations are provided, at least in part, from a memory address location associated with the index.

Abstract: A neuromorphic device may include a pre-synaptic neuron, a row line extending from the pre-synaptic neuron in a row direction, a post-synaptic neuron, a column line extending from the post-synaptic neuron in a column direction, and a synapse coupled between the row line and the column line. The synapse may be disposed in an intersection region between the row line and the column line. The post-synaptic neuron may include a subtracting circuit.

Abstract: Methods and systems are disclosed using improved Convolutional Neural Networks (CNN) for image processing. In one example, an input image is down-sampled into smaller images with a smaller resolution than the input image. The down-sampled smaller images are processed by a CNN having a last layer with a reduced number of nodes than a last layer of a full CNN used to process the input image at a full resolution. A result is outputted based on the processed down-sampled smaller images by the CNN having a last layer with a reduced number of nodes. In another example, shallow CNN networks are built randomly. The randomly built shallow CNN networks are combined to imitate a trained deep neural network (DNN).

Abstract: An information processing method including the following executed using a computer: obtaining a neural network model that solves a regression problem; obtaining input data and label data corresponding to the input data; compressing a network of the neural network model to obtain a compressed model; and changing the label data and the number of nodes in the neural network model, based on information indicating performance of the compressed model, the number of nodes being assigned to the regression problem, the information being calculated using the label data and output data which is obtained by inputting the input data to the compressed model.

Abstract: A deep learning based compression (DLBC) system applies trained models to compress binary code of an input image to a target codelength. For a set of binary codes representing the quantized coefficents of an input image, the DLBC system applies a first model that is trained to predict feature probabilities based on the context of each bit of the binary codes. The DLBC system compresses the binary code via adaptive arithmetic coding based on the determined probability of each bit. The compressed binary code represents a balance between a reconstruction quality of a reconstruction of the input image and a target compression ratio of the compressed binary code.

Abstract: Optimized learning settings of neural networks are efficiently determined by an apparatus including a processor and one or more computer readable mediums collectively including instructions that, when executed by the processor, cause the processor to train a first neural network with a learning setting; extract tentative weight data from the first neural network with the learning setting; calculate an evaluation value of the first neural network with the learning setting; and generate a predictive model for predicting an evaluation value of a second neural network with a new setting based on tentative weight data of the second neural network by using a relationship between the tentative weight data of the first neural network and the evaluation value of the first neural network.

Abstract: A system combines inputs from human processing and machine processing, and employs machine learning to improve processing of individual tasks based on comparison of human processing results. Once performance of a particular task by machine processing reaches a threshold, the level of human processing used on that task is reduced.

Abstract: A resolution improvement system includes a server, for performing, in response to a user device side request, a service for transmitting requested video data to a user device. A universal neural network file required for operation of an artificial neural network algorithm for improving the resolution of image information on the basis of the retained video data is generated, and the low-quality video data in which the generated universal neural network file and the resolution are changed to less than or equal to a preset level is transmitted to the user device. A user device performs a calculation operation on the basis of an artificial neural network algorithm that applies the received universal neural network file to the low quality video data received from the server, improving the resolution of the low quality video data according to the calculation operation, and playing back the video data with improved resolution.

Abstract: A computing device classifies unclassified observations. A first batch of noise observations is generated. (A) A first batch of unclassified observations is selected. (B) A first batch of classified observations is selected. (C) A discriminator neural network model trained to classify unclassified observations and noise observations is updated with observations that include the first batch of unclassified observations, the first batch of classified observations, and the first batch of noise observations. (D) A discriminator loss value is computed that includes an adversarial loss term computed using a predefined transition matrix. (E) A second batch of unclassified observations is selected. (F) A second batch of noise observations is generated. (G) A generator neural network model trained to generate a fake observation vector for the second batch of noise observations is updated with the second batch of unclassified observations and the second batch of noise observations. (H) (A) to (G) is repeated.

Abstract: A machine learning device includes: a state observer that observes, as a state variable, at least one of driving noise of a synchronous rotating machine and an error with respect to a preset position of a rotational position of the synchronous rotating machine determined by driving voltage; and a learner that determines a value of a parameter on a basis of the state variable observed by the state observer. The learner includes: a reward calculator that calculates a reward on a basis of the state variable; and a function updater that determines the value of the parameter on a basis of the reward that has been calculated. The reward calculator increases the reward when the driving noise is smaller than a target value of driving noise.

Abstract: Embodiments of the present disclosure disclose a method and a device for classifying questions based on artificial intelligence. The method includes: acquiring text content of a question input by a user, and performing a word segmentation process on the text content to obtain a plurality of segmentations; acquiring hidden representation vectors of the plurality of segmentations; generating a first vector of the text content according to the hidden representation vectors; and determining a target responder corresponding to the question according to the first vector and a preset classification model, and appointing the target responder to the user. The method may simplify operation steps, reduce interactions between a user and a service center, and improve efficiency of the service center.

Abstract: The boundary search test support device includes: a storage device that holds a plurality of input vectors; and an arithmetic device that executes a test by sequentially inputting the input vectors to a program generated by a neural network and acquiring output vectors which are test results, respectively generates, in a coordinate system which takes each of a predetermined plurality of elements among elements constituting the output vectors as a coordinate axis, a straight line in which the plurality of elements has a same value and a hyperplane in which a sum of values of the plurality of elements is taken as a predetermined value, and arranges a most antagonistic point and boundary vectors whose values of the elements rank higher than or equal to a predetermined ranking among the output vectors in the coordinate system, and outputs the coordinate system together with input vectors corresponding to the boundary vectors.

Abstract: Provided is a spiking neural network system for dynamical control of flexible, stable, and hybrid memory storage. An information storage method may include converting input information to a temporal pattern in a form of a spike; and storing the information that is converted to the temporal pattern in a spiking neural network. The storing may comprise storing information by applying, to the spiking neural network, a spike-timing-dependent plasticity (STDP) learning rate that is an unsupervised learning rule.

Abstract: Aspects of the present disclosure involve, a customizable system and infrastructure which can receive privacy data from varying data sources for privacy scanning, containment, and reporting. In one embodiment, data received is scanned for privacy data extraction using various data connectors and decryption techniques. In another embodiment, the data extracted is transferred to a privacy scanning container where the data is analyzed by various deep learning models for the correct classification of the data. In some instances, the data extracted may be unstructured data deriving form emails, case memos, surveys, social media posts, and the like. Once the data is classified, the data may be stored or contained according to the classification of the data. Still in another embodiment, the classified data may be retrieved by an analytics container for use in reporting.

Abstract: An emitter identification system arranged to: receive a detected signal including one or more emitter signals from one or more emitters respectively where each of the emitter signals includes a unique signal characteristic related to a unique physical feature of a hardware structure associated with each of the emitters; apply a modulation signal to the detected signal to generate pulse in-phase and quadrature (IQ) data associated with the one or more emitter signals; extract one or more amplitude envelopes associated with the one or more emitter signals, where each amplitude envelope is related to the unique signal characteristic associated with each of the one or more emitters; estimate the unique signal characteristic of each of the one or more emitter signals; estimate a number of clusters related to a number of emitter signals; and identify each of the emitters by applying an unsupervised learning function.

Abstract: In general, techniques are described for providing diversity in simulation datasets during modeling. A device comprising a memory and a processor may be configured to perform the techniques. The memory may store simulation configuration files for conducting simulations of the network device within a test environment. The processor may conduct, based on the simulation configuration files, each of the simulations with respect to the network device to collect corresponding simulation datasets indicative of an operating state of the network device. The processor may determine a level of similarity between the simulation datasets, and select, responsive to a comparison of the level of similarity to a diversity threshold, a subset of the simulation datasets. The processor may generate, based on the selected subset of the simulation datasets, a model representative of the network device that predicts, responsive to configuration parameters for the network device, an operating state of the network device.

Abstract: Disclosed are various embodiments of memristive devices comprising a number of nodes. Memristive fibers are used to form conductive and memristive paths in the devices. Each memristive fiber may couple one or more nodes to one or more other nodes. In one case, a memristive device includes a first node, a second node, and a memristive fiber. The memristive fiber includes a conductive core and a memristive shell surrounding at least a portion of the conductive core along at least a portion of the memristive fiber. The memristive fiber couples the first node to the second node through a portion of the memristive shell and at least a portion of the conductive core.

Abstract: A device, system, and method for training or prediction of a neural network. A current value may be stored for each of a plurality of synapses or filters in the neural network. A historical metric of activity may be independently determined for each individual or group of the synapses or filters during one or more past iterations. A plurality of partial activations of the neural network may be iteratively executed. Each partial-activation iteration may activate a subset of the plurality of synapses or filters in the neural network. Each individual or group of synapses or filters may be activated in a portion of a total number of iterations proportional to the historical metric of activity independently determined for that individual or group of synapses or filters. Training or prediction of the neural network may be performed based on the plurality of partial activations of the neural network.

Abstract: A normalization method for machine learning and an apparatus thereof are provided. The normalization method according to some embodiments of the present disclosure may calculate a value of a normalization parameter for an input image through a normalization model before inputting the input image to a target model and normalize the input image using the calculated value of the normalization parameter. Because the normalization model is updated based on a prediction loss of the target model, the input image can be normalized to an image suitable for a target task, so that stability of the learning and performance of the target model can be improved.

Abstract: A form filling application is configured to minimize the form filling effort of a user. The configuration follows concepts from reinforcement learning, and includes optimizing a policy for selecting agent actions in a manner that maximizes a reward signal. In the context of the form filling application, an agent action may specify one or more slots of the form for the user to fill, and further specify one or more user interfaces for filling the specified one or more slots. The reward signal may be defined as an inverse function of the user effort, so that maximizing the reward signal has the desired effect of minimizing the user effort.

Abstract: An information processing method including the following executed using a computer: obtaining a neural network model that solves a regression problem; obtaining input data and label data corresponding to the input data; compressing a network of the neural network model to obtain a compressed model; transforming the regression problem to be solved by the neural network model into a classification problem, based on information indicating performance of the compressed model, the information being calculated using the label data and output data which is obtained by inputting the input data to the compressed model; and changing a network configuration of the neural network model and transforming the label data, in accordance with the transformation from the regression problem to the classification problem.

Abstract: Systems, apparatuses, and methods for representing words or phrases, and using the representation to perform NLP and NLU tasks, where these tasks include sentiment analysis, question answering, and conference resolution. Embodiments introduce a type of deep contextualized word representation that models both complex characteristics of word use, and how these uses vary across linguistic contexts. The word vectors are learned functions of the internal states of a deep bidirectional language model (biLM), which is pre-trained on a large text corpus. These representations can be added to existing task models and significantly improve the state of the art across challenging NLP problems, including question answering, textual entailment and sentiment analysis.

Abstract: Systems and methods for generating synthetic intercorrelated data are disclosed. For example, a system may include at least one memory storing instructions and one or more processors configured to execute the instructions to perform operations. The operations may include training a parent model by iteratively performing steps. The steps may include generating, using the parent model, first latent-space data and second latent-space data. The steps may include generating, using a first child model, first synthetic data based on the first latent-space data, and generating, using a second child model, second synthetic data based on the second latent-space data. The steps may include comparing the first synthetic data and second synthetic data to training data. The steps may include adjusting a parameter of the parent model based on the comparison or terminating training of the parent model based on the comparison.

Abstract: Disclosed embodiments provide for vehicle manipulation using convolutional image processing. The convolutional image processing is accomplished using a computer, where the computer can include a multilayered analysis engine. The multilayered analysis engine can include a convolutional neural network (CNN). The computer is initialized for convolutional processing. A plurality of images is obtained using an imaging device within a first vehicle. A multilayered analysis engine is trained using the plurality of images. The multilayered analysis engine includes multiple layers that include convolutional layers hidden layers. The multilayered analysis engine is used for cognitive state analysis. The evaluating provides a cognitive state analysis. Further images are analyzed using the multilayered analysis engine. The further images include facial image data from one or more persons present in a second vehicle. Voice data is collected to augment the cognitive state analysis.

Abstract: A deep neural network architecture comprises a stack of strata in which each stratum has its individual input and an individual objective, in addition to being activated from the system input through lower strata in the stack and receiving back propagation training from the system objective back propagated through higher strata in the stack of strata. The individual objective for a stratum may comprise an individualized target objective designed to achieve diversity among the strata. Each stratum may have a stratum support subnetwork with various specialized subnetworks. These specialized subnetworks may comprise a linear subnetwork to facilitate communication across strata and various specialized subnetworks that help encode features in a more compact way, not only to facilitate communication across strata but also to increase interpretability for human users and to facilitate communication with other machine learning systems.

Abstract: Systems, methods, media, and other such embodiments described herein relate to computational analysis of data associated using tensor networks. One embodiment involves accessing a set of input data (e.g. text, images, audio, or other data associated with language or “meaning” correlations) from a memory of a computing device, and processing the input data to generate a plurality of data elements as a sequence of vectors representing the input data. This sequence of vectors is then input into a tensor network comprising a plurality of interconnected nodes, with each node comprising an operator having an associated operator value, with the operator being configured to act on a product of a vector space associated with a data element of the plurality of data elements. The tensor network outputs one or more values, with each value associated with at least one data element of set of input data.

Abstract: Systems, methods, and apparatus, including computer programs encoded on a computer storage medium, related to associative long short-term memory (LSTM) neural network layers configured to maintain N copies of an internal state for the associative LSTM layer, N being an integer greater than one. In one aspect, a system includes a recurrent neural network including an associative LSTM layer, wherein the associative LSTM layer is configured to, for each time step, receive a layer input, update each of the N copies of the internal state using the layer input for the time step and a layer output generated by the associative LSTM layer for a preceding time step, and generate a layer output for the time step using the N updated copies of the internal state.

Abstract: Methods and systems are provided for operating a neuromorphic system for generating neuron and synapse activities. The method includes: preparing at least one digital timer in the neuromorphic system, each of the at least one digital timers including multi-bit digital values; generating time signals using the at least one digital timer; emulating an analog waveform of a neuron spike; updating parameters of the neuromorphic system using the time signals and the current values of the parameters; presetting, using a processor, the digital values of the at least one digital timer to initial values when the spike input is provided to the node; and updating, using the processor, the digital values of the at least one digital timer with a specified amount when there is an absence of a spike input to the node.

Abstract: Aspects for vector operations in neural network are described herein. The aspects may include a vector caching unit configured to store a first vector and a second vector, wherein the first vector includes one or more first elements and the second vector includes one or more second elements. The aspects may further include one or more adders and a combiner. The one or more adders may be configured to respectively add each of the first elements to a corresponding one of the second elements to generate one or more addition results. The combiner may be configured to combine a combiner configured to combine the one or more addition results into an output vector.

Abstract: An apparatus and method are described for distributed and cooperative computation in artificial neural networks. For example, one embodiment of an apparatus comprises: an input/output (I/O) interface; a plurality of processing units communicatively coupled to the I/O interface to receive data for input neurons and synaptic weights associated with each of the input neurons, each of the plurality of processing units to process at least a portion of the data for the input neurons and synaptic weights to generate partial results; and an interconnect communicatively coupling the plurality of processing units, each of the processing units to share the partial results with one or more other processing units over the interconnect, the other processing units using the partial results to generate additional partial results or final results. The processing units may share data including input neurons and weights over the shared input bus.

Abstract: A multiplexed neural core circuit according to one embodiment comprises, for an integer multiplexing factor T that is greater than zero, T sets of electronic neurons, T sets of electronic axons, where each set of the T sets of electronic axons corresponds to one of the T sets of electronic neurons, and a synaptic interconnection network comprising a plurality of electronic synapses that each interconnect a single electronic axon to a single electronic neuron, where the interconnection network interconnects each set of the T sets of electronic axons to its corresponding set of electronic neurons.

Abstract: Embodiments of the invention provide a method for mapping a bipartite graph onto a neuromorphic architecture comprising of a plurality of interconnected neuromorphic core circuits. The graph includes a set of source nodes and a set of target nodes. The method comprises, for each source node, creating a corresponding splitter construct configured to duplicate input. Each splitter construct comprises a first portion of a core circuit. The method further comprises, for each target node, creating a corresponding merger construct configured to combine input. Each merger construct comprises a second portion of a core circuit. Source nodes and target nodes are connected based on a permutation of an interconnect network interconnecting the core circuits.

Abstract: A neural network conversion method and a recognition apparatus that implements the method are provided. A method of converting an analog neural network (ANN) to a spiking neural network (SNN) normalizes first parameters of a trained ANN based on a reference activation that is set to be proximate to a maximum activation of artificial neurons included in the ANN, and determines second parameters of an SNN based on the normalized first parameters.

Abstract: A system configured to identify a target in a synthetic aperture radar signal includes: a feature extractor configured to extract a plurality of features from the synthetic aperture radar signal; a spiking neural network configured to encode the features as a plurality of spiking signals; a readout neural layer configured to compute a signal identifier based on the spiking signals; and an output configured to output the signal identifier, the signal identifier identifying the target.

Abstract: A system and method to use deep learning for super resolution in a radar system include obtaining first-resolution time samples from reflections based on transmissions by a first-resolution radar system of multiple frequency-modulated signals. The first-resolution radar system includes multiple transmit elements and multiple receive elements. The method also includes reducing resolution of the first-resolution time samples to obtain second-resolution time samples, implementing a matched filter on the first-resolution time samples to obtain a first-resolution data cube and on the second-resolution time samples to obtain a second-resolution data cube, processing the second-resolution data cube with a neural network to obtain a third-resolution data cube, and training the neural network based on a first loss obtained by comparing the first-resolution data cube with the third-resolution data cube. The neural network is used with a second-resolution radar system to detect one or more objects.

Abstract: A method of operating a polishing system includes training a plurality of models using a machine learning algorithm to generate a plurality of trained models, each trained model configured to determine a characteristic value of a layer of a substrate based on a monitoring signal from an in-situ monitoring system of a semiconductor processing system, storing the plurality of trained models, receiving data indicating a characteristic of a substrate to be processed, selecting one of the plurality of trained models based on the data, and passing the selected trained model to the processing system.

Abstract: A multi-stream cross correlator for spiking neural networks, where each stream contains significant stochastic content. At least one event occurs, with a fixed temporal relationship across at least two streams. Each stream is treated as a Frame Of Reference (FOR), and subject to an adjustable delay based on comparison to the Other streams. For each spike of the FOR, a timing analysis, relative to the last and current FOR spikes, is completed by comparing Post and Pre accumulators. Also, a new timing analysis is begun, with the current FOR spike, by restarting the production of Post and Pre weighting functions, the values of which are accumulated, upon the occurrence of each Other spike, until a next FOR spike. A one-spike delay unit can be used, if time-neutral conflict resolution is used. The average spike rate of the FOR can be determined and used for the Post and Pre weighting functions.

Abstract: A method and system providing a multi-memristive synaptic element for a cognitive computing system. The multi-memristive synaptic element comprises an array of memristive devices. The method comprises arbitrating a synaptic weight allocation, a related synaptic weight being represented by a synaptic weight variable of said multi-memristive synaptic element, updating said synaptic weight variable by a delta amount, and assigning said memristive devices to elements of a clock-like ordered circular list for selecting a particular memristor of said memristive devices requiring to be updated by a deterministic, periodic global clock that points to a different memristor at every clock tick, such that said multi-memristive synaptic element has a larger dynamic range and a more linear conductance response than a single memristor synaptic element.

Abstract: A control system for an operable system such as a flow control system or temperature control system. The system operates in a control loop to regularly update a model with respect at least one optimizable input variable based on the detected variables. The model provides prediction of use of the input variables in all possible operation points or paths of the system variables which achieve an output setpoint. In some example embodiments, the control loop is performed during initial setup and subsequent operation of the one or more operable elements in the operable system. The control system is self-learning in that at least some of the initial and subsequent parameters of the system are determined automatically during runtime.

Abstract: Some examples relate to collection of performance metrics from a device. In an example, performance metrics for collection from a first device may be selected. The performance metrics may be indexed by assigning an index entry to respective performance metrics on the first device. A fixed sequence of the performance metrics may be maintained on first device. The fixed sequence of the performance metrics along with the index entry assigned to the respective performance metrics may be shared with a second device. A first performance data of the respective performance metrics on first device may be determined. The first performance data of the respective performance metrics may be shared with second device. The sharing may comprise sending, to second device, the index entry and the first performance data of the respective performance metrics in an order corresponding to the fixed sequence of the performance metrics on first device.

Abstract: A resistive memory cell is connected in circuitry which has a first input terminal for applying neuron input signals including a read portion and a write portion. The circuitry includes a read circuit producing a read signal dependent on resistance of the memory cell, and an output terminal providing a neuron output signal, dependent on the read signal in a first state of the memory cell. The circuitry also includes a storage circuit storing a measurement signal dependent on the read signal, and a switch set operable to supply the read signal to the storage circuit during application of the read portion of each neuron input signal to the memory cell, and, after application of the read portion, to apply the measurement signal in the apparatus to enable resetting of the memory cell to a second state.

Abstract: Disclosed are implementations that include a method comprising applying at least one input signal to a power amplification system, that includes a transmit chain with a power amplifier (PA) producing output with non-linear distortions, to produce at least one output signal, and measuring at least one observed signal of the output signal using an observation receiver coupled to an output of the transmit chain, the observation receiver having a receive bandwidth smaller than a transmit chain bandwidth of the transmit chain. Measuring the at least one observed signal includes measuring multiple frequency segments of output signal. The method further includes determining one or more sets of digital predistortion coefficients based on the measured multiple frequency segments of the at least one output signal, with each of the sets of digital predistortion coefficients being associated with a respective set of operating conditions of the power amplification system.

Abstract: Systems and methods are described for quantifying the response of a biological system to one or more perturbations. First and second datasets corresponding to a response of a biological system to first and second treatments are received. A plurality of computational network models that represent the biological system are provided, each model including nodes representing a plurality of biological entities and edges representing relationships between the nodes in the model. A first set of scores is generated, representing the perturbation of the biological system based on the first dataset and the plurality of models, and a second set of scores representing the perturbation of the biological system based on the second dataset and the plurality of computational models. One or more biological impact factors are generated based on each of the first set and second set of scores that represent the biological impact of the perturbation on the biological system.

Abstract: An abnormality detection device of an embodiment includes an encoder, a first identifier, a decoder, and a second identifier. The encoder is configured to compress input data using a compression parameter to generate a compressed data. The first identifier is configured to determine whether a distribution of the compressed data input by the encoder is a distribution of the compressed data or a prior distribution prepared in advance, and inputs a first identification result to the encoder. The decoder is configured to decode the compressed data using a compressing parameter to generate reconstructed data. The second identifier is configured to determine whether the reconstructed data input by the decoder is the reconstruction data or the input data and outputs a second identification result to the encoder and the decoder.

Abstract: A sequence of images depicting an object is captured, e.g., by a camera at a point-of-sale terminal in a retail store. The object is identified, such as by a barcode or watermark that is detected from one or more of the images. Once the object's identity is known, such information is used in training a classifier (e.g., a machine learning system) to recognize the object from others of the captured images, including images that may be degraded by blur, inferior lighting, etc. In another arrangement, such degraded images are processed to identify feature points useful in fingerprint-based identification of the object. Feature points extracted from such degraded imagery aid in fingerprint-based recognition of objects under real life circumstances, as contrasted with feature points extracted from pristine imagery (e.g., digital files containing label artwork for such objects).

Abstract: A neuromorphic device is provided. The neuromorphic device may include a pre-synaptic neuron, a synapse electrically connected with the pre-synaptic neuron through a row line, and a post-synaptic neuron electrically connected with the synapse through a column line. The post-synaptic neuron may include a post-neuron circuit and a post-neuron transfer function circuit electrically connected to the column line. The post-neuron transfer function circuit may include a first inverting circuit including at least one first pull-up transistor and at least two first pull-down transistors, the pull-down transistors being electrically connected with each other in parallel.

Abstract: Systems and methods for classifying binary data based training data having a predefined sample size is obtained. The training data is composed of separable binary datasets. An exact bound on Vapnik-Chervonenkis (VC) dimension of a classifier for the training data is determined. The exact bound is based one or more variables defining the hyperplane. The exact bound may be minimized for generating a classifier for predicting one class to which a given data sample of the training data belongs.

Abstract: A method and system are provided for updating synapse weight values in neuromorphic system with Spike Time Dependent Plasticity model. The method includes selectively performing, by a hardware-based synapse weight incrementer or decrementer, one of a synapse weight increment function or decrement function, each using a respective lookup table, to generate updated synapse weight values responsive to spike timing data. The method further includes storing the updated synapse weight values in a memory. The method additionally includes performing, by a hardware-based processor, a learning process to integrate the updated synapse weight values stored in the memory into the Spike Time Dependent Plasticity model neuromorphic system for improved neuromorphic simulation.

Abstract: A method for load balancing in a computer network includes receiving application information for an application and information relating to an artificial neural network (NN) computation to be executed by the application. A configuration is derived for one or more network devices based on the application information and the information relating to the NN computation. The configuration is installed in the one or more network devices such that at least one of the network devices on a path of a network packet performs a subset of the NN computation and encodes a result of the subset of the NN computation into a header of the network packet.

Abstract: An analog neuromorphic circuit is disclosed having resistive memories that provide a resistance to an input voltage signal as the input voltage signal propagates through the resistive memories generating a first output voltage signal and to provide a resistance to a first error signal that propagates through the resistive memories generating a second output voltage signal. A comparator generates the first error signal that is representative of a difference between the first output voltage signal and the desired output signal and generates the first error signal so that the first error signal propagates back through the plurality of resistive memories. A resistance adjuster adjusts a resistance value associated with each resistive memory based on the first error signal and the second output voltage signal to decrease the difference between the first output voltage signal and the desired output signal.