Abstract: Techniques for characterizing an optical system (for example, a printer verifier) are provided. In this regard, the optical system may be characterized for scanning a printed image. The characterization of the optical system includes determining an effective aperture size of the optical system, and correspondingly an effective resolution at which the optical system can be configured to scan a portion of the printed image according to verification requirements.

Abstract: An optical scanner device includes at least one image capture device and a transmitter of at least one aimer beam. The scanner device determines ranging to a subject using the at least one aimer beam projected to reflect off of a surface of the subject, and detects a position of the aimer-beam reflection within an image frame captured by the image-capture device, the position being a primary indicator of a distance to the subject from the optical scanner device. A secondary indicator of the distance to the subject within the image frame in combination with the first indicator is used to help detect the aimer beam reflection against noise and detect an occurrence of an optical misalignment with possible self-correction of calibration after such misalignment.

Abstract: The present invention relates to a chip counter, which transmits an X-ray beam through a tape reel around which a tape having a plurality of semiconductor chips mounted in a row therein is wound, acquires an image scattered or diffracted by the semiconductor chips, and processes the acquired image, so as to count the number of the semiconductor chips, wherein: the X-ray beam transmitted through the tape reel (1) is sensed by a fluorescent intensifying screen (60); a fluorescent light emitted from the fluorescent intensifying screen (60) according to the sensing of the X-ray beam is captured by a camera (70), so that the number of the semiconductor chips is counted from an image in which the semiconductor chips are displayed by a dotted image; and the camera (70) is protected by an X-ray beam shielding member (100: 110; 120; and 130).

Abstract: In the context of gate-model quantum computing, atoms (or polyatomic molecules) are excited to respective Rydberg states to foster intra-gate interactions. Rydberg states with relatively high principal quantum numbers are used for relatively distant intra-gate interactions and require relatively great inter-gate separations to avoid error-inducing inter-gate interactions. Rydberg states with relatively low principal quantum numbers can be used for intra-gate interactions over relatively short intra-gate distances and require relatively small inter-gate separations to avoid error-inducing inter-gate interactions. The relatively small inter-gate separations provide opportunities for parallel gate executions, which, in turn, can provide for faster execution of the quantum circuit constituted by the gates.

Abstract: This disclosure describes a quantum noise process analysis method, device, and storage medium, in the field of quantum processing technologies. The method may include performing quantum process tomography (QPT) on a quantum noise process of a target quantum system, to obtain dynamical maps of the quantum noise process, wherein the QPT involves at least one measurement of the target quantum. The method further includes extracting transfer tensor maps (TTMs) of the quantum noise process from the dynamical maps; and analyzing the quantum noise process according to the TTMs. The TTM is used for representing a dynamical evolution of the quantum noise process to reflect the law of evolution of the dynamical maps of the quantum noise process over time.

Abstract: In a general aspect, information is encoded in data qubits in a three-dimensional device lattice. The data qubits reside in multiple layers of the three-dimensional device lattice, and each layer includes a respective two-dimensional device lattice. A three-dimensional color code is applied in the three-dimensional device lattice to detect errors in the data qubits residing in the multiple layers. A two-dimensional color code is applied in the two-dimensional device lattice in each respective layer to detect errors in one or more of the data qubits residing in the respective layer.

Abstract: A method for providing teleportation services includes receiving, by a computing device, a first signal. The first signal indicates a request for a teleportation event between a first quantum computing system (QCS) and a second QCS. A first set of qubits is associated with the first QCS. A second set of qubits is associated with the second QCS. In response to receiving the first signal, the computing device causes an allocation of a first qubit of the first set of qubits for the teleportation event. In response to receiving the signal, the computing device causes an allocation of a second qubit of the second set of qubits for the teleportation event. The computing device receives a second signal that indicates a successful completion of the teleportation event. In response to receiving the second signal, the computing system causes a deallocation of the first qubit of the first set of qubits.

Abstract: In one embodiment, a device classification service receives telemetry data indicative of behavioral characteristics of a plurality of devices in a network. The service obtains side information for the telemetry data. The service applies metric learning to the telemetry data and side information, to construct a distance function. The service uses the distance function to cluster the telemetry data into device clusters. The service associates a device type label with a particular device cluster.

Abstract: Techniques for data integration and labeling are provided. Training real-world signal data is collected for a physical environment, where the training real-world signal data comprises at least one of (i) coordinate information or (ii) a direction to move. Simulated signal data is generated for a first portion of the physical environment, and an aggregate data set is generated comprising the training real-world signal data and the simulated signal data. A machine learning (ML) model is trained using the aggregate data set. A first real-world data point is received, where the first real-world data point does not include coordinate information, and the first real-world data point is labeled based at least in part on coordinate information of the aggregate data set.

Abstract: Systems, methods, and computer program products are described herein for managing a lifecycle of a machine learning (ML) application from a provider point of view. Within a data intelligence platform, a package having ML scenarios and a training pipeline is generated. The training pipeline includes training logic associated with a defined workflow for training the ML application. The data intelligence platform is synchronized with a first database via an application programming interface. The first database generates a transport request containing the package. The transport request facilitates publication of content from the ML application. The ML application is assembled from the transport request within a second database. ML content is displayed on a graphical user interface associated with the second database.

Abstract: A method includes identifying records in a database for labeling; presenting one of the records in the database to a first labeling entity; and receiving a first observation on an information source in the one of the records from the first labeling entity. The first observation has one of a plurality of observation types associated therewith. The plurality of observation types including a validation observation type in which the first observation comprises a confirmation of whether a second observation on the information source in the one of the records from another labeling entity is accurate and an edit for the second observation when the second observation is confirmed as inaccurate. The one of the records is updated in the database with the first observation on the information source in the one of the records from the first labeling entity.

Abstract: AI-assisted Connected Home systems for improving power efficiency at homes and offices are described. The system may perform operations including: receiving, from each of a plurality of power receptacles, power usage information of an electrical device attached to the power receptacle; determining, for each of the plurality of power receptacles, a plurality of power usage metrics of the power receptacle based on the power usage information; feeding the plurality of power usage metrics into a machine learning model to obtain a priority of the electrical device attached to the power receptacle, wherein the priority is one of a plurality of pre-configured priorities; obtaining a power management signal; and transmitting a plurality of control signals to the plurality of power receptacles based on the power management signal and respective priorities of the plurality of electrical devices attached to the plurality of power receptacles.

Abstract: Systems and methods for labeling data are disclosed. An example method may be performed by one or more processors of a labeling system and include retrieving labeled data, identifying characteristics predictive of labels that would be entered for unlabeled data items having the respective characteristics based on the labeled data, training an analysis model to predict labels that would be entered for unlabeled data items, generating, for unlabeled data items, using the trained analysis model, a prediction of a label that will be entered for the respective unlabeled data item if the respective unlabeled data item is presented for labeling, selecting, based on the generated predictions, a subset of unlabeled data items to be presented for labeling, receiving labels for the subset of unlabeled data items, determining that a completion criteria associated with the trained analysis model is met, and generating labels for remaining unlabeled data items.

Abstract: Disclosed are systems, methods, and non-transitory computer-readable media for sharing, on a distributed database, a database application to a first user of the distributed database, the database application generated by a second user of the distributed database. The training dataset includes a first database training dataset from the first user of the distributed database and a second database training dataset from the second user of the distributed database, the first database training dataset and the second database training dataset including non-overlapping dataset features. The database application further identifies a query from the second user to train the machine learning model on the training dataset and generates a trained machine learning model by training the machine learning model on a joined dataset according to the query. The database application generates outputs from the trained machine learning model by applying the trained machine learning model on new data.

Abstract: This disclosure describes systems and methods for using an estimator to produce values for dependent variables of streaming objects based on values of independent variables of the objects. The systems and methods may include continuously tuning the estimator based on any objects received with pre-populated values for the dependent variables.

Abstract: An apparatus and methods for training an educational machine-learning model, the apparatus includes a sensory device configured to capture an external datum pertaining to a user, at least a processor in communication with the sensory device, and a memory containing instructions configuring the at least a processor to receive the user data, wherein the user data includes the external datum captured by the sensory device, and a user input accepted through a visual interface, authenticate the user as a function of the external datum using a user authentication module, generate educational training data as a function of the user data, train an educational machine-learning model using the educational training data, and determine a user input modifier as a function of the trained educational machine-learning model.

Abstract: Methods and systems are presented for generating a machine learning model using enhanced gradient boosting techniques. The machine learning model is configured to receive inputs corresponding to a set of features and to produce an output based on the inputs. The machine learning model includes multiple layers, wherein each layer includes multiple models. To generate the machine learning model, multiple models are built and trained in parallel for each layer of the machine learning model. The multiple models use different subsets of features to produce corresponding output values. After a layer in built and trained, a collective error may be determined for the layer based on the output values from the different models in the layer. An additional layer of models may be added to the machine learning model to reduce the collective error of a previous layer.

Abstract: Systems and methods for training models to improve fairness.

Abstract: Input data associated with a machine learning model is classified into a plurality of clusters. A plurality of linear surrogate models are generated. One of the plurality of linear surrogate models corresponds to one of the plurality of clusters. A linear surrogate model is configured to output a corresponding prediction based on input data associated with a corresponding cluster. Prediction data associated with the machine learning model and prediction data associated with the plurality of linear surrogate models are outputted.

Abstract: Apparatuses, systems, and techniques to identify a goal of a demonstration. In at least one embodiment, video data of a demonstration is analyzed to identify a goal. Object trajectories identified in the video data are analyzed with respect to a task predicate satisfied by a respective object trajectory, and with respect to motion predicate. Analysis of the trajectory with respect to the motion predicate is used to assess intentionality of a trajectory with respect to the goal.

Abstract: Embodiments of the present disclosure include systems and methods for training transformer models using position masking. In some embodiments, a set of data for training a transformer model is received. The set of data includes a sequence of tokens and a set of position values. Each position value in the set of position values represents a position of a token in the sequence of tokens relative to other tokens in the sequence of tokens. A subset of the set of position values in the set of data is selected. Each position value in the subset of the set of position values is replaced with a second defined value to form a second set of defined values. The transformer model is trained using the set of data.

Abstract: Techniques for neural network processing using specialized data representation are disclosed. Input data for manipulation in a layer of a neural network is obtained. The input data includes image data, where the image data is represented in bfloat16 format without loss of precision. The manipulation of the input data is performed on a processor that supports single-precision operations. The input data is converted to a 16-bit reduced floating-point representation, where the reduced floating-point representation comprises an alternative single-precision data representation mode. The input data is manipulated with one or more 16-bit reduced floating-point data elements. The manipulation includes a multiply and add-accumulate operation. The manipulation further includes a unary operation, a binary operation, or a conversion operation. A result of the manipulating is forwarded to a next layer of the neural network.

Abstract: In one implementation, a method is implemented by a neural network device and includes inputting a representation of topological structures in patterns of activity in a source neural network, wherein the activity is responsive to an input into the source neural network, processing the representation, and outputting a result of the processing of the representation. The processing is consistent with a training of the neural network to process different such representations of topological structures in patterns of activity in the source neural network.

Abstract: Disclosed is an accuracy compensation method for discharge caustic alkali concentration measuring device in evaporation process, comprising following steps: step 1. collecting process data of instrument values and laboratory values of alkali liquor refractive index, temperature and caustic alkali concentration in the evaporation process; step 2. performing sliding average filtering, time series matching and normalization on the process data collected in step 1 to obtain preprocessed process data; step 3. inputting the preprocessed process data into an accuracy compensation model of the caustic alkali concentration measuring device to obtain compensation values; step 4. adding the compensation values of the caustic alkali concentration to the instrument values to realize on-line compensation of the caustic alkali concentration.

Abstract: A method for model adaptation, an electronic device, and a computer program product are disclosed. For example, the method comprises processing first input data by using a first machine learning model having first parameter set values, to obtain first feature information of the first input data, the first machine learning model having a capability of self-ordering and the first parameter set values being updated after the processing of the first input data; generating a first classification result for the first input data based on the first feature information by using a second machine learning model having second parameter set values; processing second input data by using the first machine learning model having the updated first parameter set values, to obtain second feature information of the second input data; and generating a second classification result for the second input data based on the second feature information by using the second machine learning model having the second parameter set values.

Abstract: A neuron circuit can switch between two functions: as an input neuron circuit, and as a hidden neuron circuit. An error circuit can switch between two functions: as a hidden error circuit, and as an output neuron circuit. A switching circuit is configured to be capable of changing the connections between the neuron circuit, a synapse circuit, and the error circuit. The synapse circuit includes an analog memory that stores data that corresponds to the connection strength between the input neuron circuit and the hidden neuron circuit or between the hidden neuron circuit and the output neuron circuit, a writing circuit that changes the data in the analog memory, and a weighting circuit that weights an input signal in reaction to the data of the analog memory and outputs the weighted output signal. The analog memory includes a transistor comprising an oxide semiconductor with extremely low off-state current.

Abstract: Neural network inference may be performed by configuration of a device including an accumulation memory, a plurality of convolution modules configured to perform mathematical operations on input values, a plurality of adder modules configured to sum values output from the plurality of convolution modules, and a plurality of convolution output interconnects connecting the plurality of convolution modules, the plurality of adder modules, and the accumulation memory. The accumulation memory is an accumulation memory allocation of a writable memory block having a reconfigurable bank width, and each bank of the accumulation memory allocation is a virtual combination of consecutive banks of the writable memory block.

Abstract: According to an embodiment, an inference system includes a recurrent neural network circuit, an inference neural network, and a control circuit. The recurrent neural network circuit receives M input signals and outputs N intermediate signals, where M is an integer of 2 or more and N is an integer of 2 or more. The inference neural network circuit receives the N intermediate signals and outputs L output signals, where L is an integer of 2 or more. The control circuit adjusts a plurality of coefficients that are set to the recurrent neural network circuit and adjusts a plurality of coefficients that are set to the inference neural network circuit. The control circuit adjusts the coefficients set to the recurrent neural network circuit according to a total delay time period from timing for applying the M input signals until timing for firing the L output signals.

Abstract: A system may comprise a neural processing unit (NPU) including at least one memory and a plurality of processing elements (PEs) capable of performing operations for at least one artificial neural network (ANN) model. The plurality of PEs may include an adder, a multiplier, and an accumulator. The plurality of PEs may include a first group of PEs configured to operate on a first portion of a clock signal and a second group of PEs configured to operate on a second portion of the clock signal.

Abstract: Numerous embodiments are disclosed for programmable output blocks for use with a VMM array within an artificial neural network. In one embodiment, the gain of an output block can be configured by a configuration signal. In another embodiment, the resolution of an ADC in the output block can be configured by a configuration signal.

Abstract: A method for realizing a Hadamard product, a device and a storage medium, includes: acquiring a plurality of to-be-treated optical signals with unequal wavelengths; inputting the to-be-treated optical signals into a wavelength division multiplexer; by using the wavelength division multiplexer, feeding the to-be-treated optical signals to a micro-ring-resonator component, wherein the micro-ring-resonator component includes a plurality of micro-ring-resonator groups each of which is formed by two micro-ring resonators with equal radii; and applying a corresponding electric current to the micro-ring-resonator component, to obtain a result of the Hadamard product according to an outputted light intensity.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for reinforcement learning with scheduled auxiliary tasks. In one aspect, a method includes maintaining data specifying parameter values for a primary policy neural network and one or more auxiliary neural networks; at each of a plurality of selection time steps during a training episode comprising a plurality of time steps: receiving an observation, selecting a current task for the selection time step using a task scheduling policy, processing an input comprising the observation using the policy neural network corresponding to the selected current task to select an action to be performed by the agent in response to the observation, and causing the agent to perform the selected action.

Abstract: Examples are disclosed that relate to the restoration of degraded images acquired via a behind-display camera. One example provides a method of training a machine learning model, the method comprising inputting training image pairs into the machine learning model, each training image pair comprising an undegraded image and a degraded image that represents an appearance of the undegraded image to a behind-display camera, and training the machine learning model using the training image pairs to generate frequency information that is missing from the degraded images.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for generating an output sequence from an input sequence. In one aspect, one of the systems includes an encoder neural network configured to receive the input sequence and generate encoded representations of the network inputs, the encoder neural network comprising a sequence of one or more encoder subnetworks, each encoder subnetwork configured to receive a respective encoder subnetwork input for each of the input positions and to generate a respective subnetwork output for each of the input positions, and each encoder subnetwork comprising: an encoder self-attention sub-layer that is configured to receive the subnetwork input for each of the input positions and, for each particular input position in the input order: apply an attention mechanism over the encoder subnetwork inputs using one or more queries derived from the encoder subnetwork input at the particular input position.

Abstract: In some embodiments, a method for optimal parallel execution of a simulation of a design is provided. A computing device extracts one or more features from the design. The computing device provides at least the one or more features as inputs to one or more machine learning models to determine one or more predictions of execution times. The computing device determines an optimum execution architecture based on the one or more predictions of execution times. The computing device distributes portions of the design for simulation based on the optimum execution architecture. In some embodiments, one or more machine learning models are trained to generate outputs for predicting an optimal parallel execution architecture for simulation of a design.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for processing inputs using a neural network system that includes a batch normalization layer. One of the methods includes receiving a respective first layer output for each training example in the batch; computing a plurality of normalization statistics for the batch from the first layer outputs; normalizing each component of each first layer output using the normalization statistics to generate a respective normalized layer output for each training example in the batch; generating a respective batch normalization layer output for each of the training examples from the normalized layer outputs; and providing the batch normalization layer output as an input to the second neural network layer.

Abstract: In one embodiment, a method includes by a computing device, detecting a sensory input, identifying, using a machine-learning model, one or more attributes associated with the machine-learning model, wherein the attributes are identified based on the sensory input in accordance with the model's training, and presenting the attributes as output. The identifying may be performed at least in part by an inference engine that interacts with the model. The sensory input may include an input image received from a camera, and the model may identify the attributes based on an input object in the input image in accordance with the model's training. The model may include a convolutional neural network trained using training data that associates training sensory input with the attributes. The training sensory input may include a training image of a training object, and the input object may be classified in the same class as the training object.

Abstract: Embodiments generate machine learning predictions to discover target device energy usage. One or more trained machine learning models configured to discover target device energy usage from source location energy usage can be stored. Multiple instances of source location energy usage over a period of time can be received for a given source location. Using the trained machine learning model, multiple discovery predictions for the received instances of source location energy usage can be generated, the discovery predictions comprising a prediction about a presence of target device energy usage within the instances of source location energy usage. And based on the multiple discovery predictions, an overall prediction about a presence of target device energy usage within the given source location's energy usage over the period of time can be generated.

Abstract: Provided are continuously learning and optimizing artificial intelligence (AI) adaptive neural network (ANN) computer modeling methods and systems, designated “human affect computer modeling” (HACM) or “affective neuron” (AN) and, more particularly, to AI methods, systems and devices, that can recognize, interpret, process and simulate human reactions and affects such as emotional responses to internal and external sensory stimuli, that provides real-time reinforcement learning modeling that reproduces human affects and/or reactions, wherein the human affect computer modeling (HACM) can be used singularly or collectively for modeling and predicting complex human reactions and affects.

Abstract: A computer-implement method comprises: training a classifier with labeled data from a dataset; classifying, by the trained classifier, unlabeled data from the dataset; providing, by the classifier to a policy gradient, a reward signal for each data/query pair; transferring, by the classifier to a ranker, learning; training, by the policy gradient, the ranker; ranking data from the dataset based on a query; and retrieving data from the ranked data in response to the query.

Abstract: One embodiment provides for a computer-readable medium storing instructions that cause one or more processors to perform operations comprising determining a per-layer scale factor to apply to tensor data associated with layers of a neural network model and converting the tensor data to converted tensor data. The tensor data may be converted from a floating point datatype to a second datatype that is an 8-bit datatype. The instructions further cause the one or more processors to generate an output tensor based on the converted tensor data and the per-layer scale factor.

Abstract: A method for determining a final architecture for a neural network to perform a particular machine learning task is described.

Abstract: A neural processing device and method for pruning thereof are provided. The neural processing device includes a processing unit configured to perform calculations, an L0 memory configured to store input and output data of the processing unit, wherein the input and output data include a two-dimensional weight matrix and a weight manipulator configured to receive the two-dimensional weight matrix and partition it into preset sizes to thereby generate partitioned matrices, to generate a pruning matrix by pruning the partitioned matrix, and to transmit the pruning matrix to the processing unit.

Abstract: In some aspects, systems and methods for efficiently clustering a large-scale dataset for improving the construction and training of machine-learning models, such as neural network models, are provided. A dataset used for training a neural network model configured can be clustered into a first set of clusters and a second set of clusters. The neural network model can be constructed with a number of nodes in a hidden layer that is based on the number of clusters in the first set of clusters. The neural network can be trained based on training samples selected from the second set of clusters. In some aspects, the trained neural network model can be utilized to satisfy risk assessment queries to compute output risk indicators for target entities. The output risk indicator can be used to control access to one or more interactive computing environments by the target entities.

Abstract: A neural network system includes a first neural network configured to predict a mean value output and epistemic uncertainty of the output given input data, and a second neural network configured to predict total uncertainty of the output of the first neural network. The second neural network is trained to predict total uncertainty of the output of the first neural network given the input data through a training process involving minimizing a cost function that involves differences between a predicted mean value of a geophysical property of a geological formation from the first neural network and a ground-truth value of the geophysical property of the geological formation. The neural network system further includes one or more processors configured to run a software module that determines aleatoric uncertainty of the output of the first neural network based on the epistemic uncertainty of the output and the total uncertainty of the output.

Abstract: A method recognizes objects in an environment of a vehicle including a sensor system having at least one sensor unit for registering the environment and an evaluation unit for evaluating sensor data provided by the at least one sensor unit. The method includes receiving the sensor data in the evaluation unit, the sensor data including a plurality of chronologically successive measurements; inputting the sensor data into a machine learning module; and outputting an object state of at least one object estimated based on the sensor data by way of the machine learning module. The method further includes determining a plurality of chronologically successive future object states based on the estimated object state; ascertaining deviations between the future object states and measurements chronologically corresponding to the future object states based on the sensor data; and correcting a machine learning algorithm of the module based on the deviations.

Abstract: A processor-implemented method of generating feature data includes: receiving an input image; generating, based on a pixel value of the input image, at least one low-bit image having a number of bits per pixel lower than a number of bits per pixel of the input image; and generating, using at least one neural network, feature data corresponding to the input image from the at least one low-bit image.

Abstract: The proposed model is a Variational Autoencoder having a learnable prior that is parametrized with a Tensor Train (VAE-TTLP). The VAE-TTLP can be used to generate new objects, such as molecules, that have specific properties and that can have specific biological activity (when a molecule). The VAE-TTLP can be trained in a way with the Tensor Train so that the provided data may omit one or more properties of the object, and still result in an object with a desired property.

Abstract: Automated development and training of deep forest models for analyzing data by growing a random forest of decision trees using data, determining Out-of-bag (OOB) predictions for the forest, appending the OOB predictions to the data set, and growing an additional forest using the data set including the appended OOB predictions, and combining the output of the additional forest, then utilizing the model to classify data outside the training data set.

Abstract: Aspects include processors configured to (or include program code that causes a processor to) provide for data classifier devices that extract from structured text business data inputs, via natural language understanding processing, training set data elements (for example, training keywords, training concepts, training entities, and/or training taxonomy classifications, etc.). The aspects identify associations within the structured training business data of each of a plurality of business class categories with respective ones of the extracted training set data elements; and build a logical relationship data classification training knowledge base ontology that connects ones of the business classes to respective associated ones of the extracted training data elements as questions, into a plurality of knowledge base ontology question-business class associations.