Abstract: For reconstruction in medical imaging using a scan protocol with repetition, a machine learning model is trained for reconstruction of an image for each repetition. Rather than using a loss for that repetition in training, the loss based on an aggregation of images reconstructed from multiple repetitions is used to train the machine learning model. This loss for reconstruction of one repetition based on aggregation of reconstructions for multiple repetitions is based on deep set-based deep learning. The resulting machine-learned model may better reconstruct an image from a given repetition and/or a combined image from multiple repetitions than a model learned from a loss per repetition.

Abstract: A neural processing unit (NPU) for testing a component during runtime is provided. The NPU may include a plurality of functional components including a first functional component and a second functional component. At least one of the plurality of functional components may be driven for calculation of an artificial neural network. Another one of the plurality of functional components may be selected as a component under test (CUT). A scan test may be performed on the at least one functional component selected as the CUT. A tester for detecting a defect of an NPU is also provided. The tester may include a component tester configured to communicate with at least one functional component of the NPU, select the at least one functional component as a CUT, and perform a scan test for the selected CUT.

Abstract: To perform a beam search operation on an input tensor using a data processor with native hardware support, the data processor can be programmed with a set of instructions. The set of instructions can include a first machine instruction that operates on the input tensor to obtain N largest values in the input tensor, a second machine instruction that operates on the input tensor to obtain indices corresponding to the N largest values in the input tensor, and a third machine instruction that operates on the input tensor to replace the N largest values in the input tensor with a minimum value.

Abstract: Variational Autoencoders (VAEs) have been shown to be effective in modeling complex data distributions. Conventional VAEs operate with fully-observed data during training. However, learning a VAE model from partially-observed data is still a problem. A modified VAE framework is proposed that can learn from partially-observed data conditioned on the fully-observed mask. A model described in various embodiments is capable of learning a proper proposal distribution based on the missing data. The framework is evaluated for both high-dimensional multimodal data and low dimensional tabular data.

Abstract: A method of training one or more neural networks, the one or more neural networks being for use in lossy image or video encoding, transmission and decoding, the method comprising the steps of: receiving an input image at a first computer system; encoding the input image using a first neural network to produce a latent representation; decoding the latent representation using a second neural network to produce an output image, wherein the output image is an approximation of the input image; evaluating a function based on a difference between the output image and the input image; updating the parameters of the first neural network and the second neural network based on the evaluated function; and repeating the above steps using a first set of input images to produce a first trained neural network and a second trained neural network; wherein the difference between the output image and the input image is determined based on the output of a neural network acting as a discriminator; the parameters of the neural netw

Abstract: This disclosure relates to techniques for performing communication in a wireless communication system. The communication may be semantic communication and/or may use a programmable protocol stack. A protocol stack at a transmitter and/or receiver may include customization from an application platform, e.g., which may replace one or more layers relative to a non-customized protocol stack. A transmitter may transmit data, via the customized protocol stack, using a best effort data channel, e.g., with transmission characteristics that are relatively lossy in comparison to a data channel used by the non-customized protocol stack. A receiver may receive the data and may, e.g., if the data is corrupted, select one or more model to recover the data and/or determine whether reception is successful, e.g., from a semantic point of view. The proposed system may support higher data rates, less retransmissions, and/or better quality of user experience.

Abstract: Systems, apparatuses, methods, and computer program products are disclosed for distillation of a natural language processing model. An example method includes receiving, by communications circuitry, a set of text data comprising a set of observations and predicting, by processing circuitry and using the NLP model, classifications for each observation in the text data. The example method further includes generating, by model training engine, a balanced sampled data structure based on the predicted classifications for each observation in the text data and training, by the model training engine, a surrogate model using the balanced sampled data structure. The example method further includes identifying, by an interpreter and from the surrogate model, a set of most-influential tokens in the text data.

Abstract: An illustrative method includes executing, by a virtual meeting platform, a virtual meeting; executing, by the virtual meeting platform and during the virtual meeting, a collaborative slideshow presentation, the executing of the collaborative slideshow presentation including accessing a collaborative slideshow file from a storage location of a file sharing platform and using the collaborative slideshow file to provide, for display within a graphical user interface view, the collaborative slideshow presentation to a plurality of participants; and facilitating, by the virtual meeting platform and during the virtual meeting based on user designated control permissions, two or more participants included in the plurality of participants having concurrent control of the collaborative slideshow presentation.

Abstract: A method for modularizing high dimensional neural networks into neural networks of lower input dimensions. The method is suited to generating full-DOF robot grasping actions based on images of parts to be picked. In one example, a first network encodes grasp positional dimensions and a second network encodes rotational dimensions. The first network is trained to predict a position at which a grasp quality is maximized for any value of the grasp rotations. The second network is trained to identify the maximum grasp quality while searching only at the position from the first network. Thus, the two networks collectively identify an optimal grasp, while each network's searching space is reduced. Many grasp positions and rotations can be evaluated in a search quantity of the sum of the evaluated positions and rotations, rather than the product. Dimensions may be separated in any suitable fashion, including three neural networks in some applications.

Abstract: Embodiments of the present disclosure provide a method, an electronic device, and a computer program product for information processing. In an information processing method, a first network state representation and a first content request event of an emulated network are provided from an emulator to an agent for reinforcement learning, wherein the first content request event indicates that a request node in the emulated network requests target content stored in a source node. The emulator receives first action information from the agent, wherein the first action information indicates a first caching action determined by the agent, the first caching action including caching the target content in at least one caching node between the request node and the source node. The emulator collects, based on the execution of the first caching action in the emulated network, first training data for training the agent.

Abstract: Implementations relate to determining a well-formed phrase to suggest to a user to submit in lieu of a not well-formed phrase. The suggestion is rendered via an interface that is provided to a client device of the user. Those implementations relate to determining that a phrase is not well-formed, identifying alternate phrases that are related to the not well-formed phrase, and scoring the alternate phrases to select one or more of the alternate phrases to render via the interface. Some of those implementations are related to identifying that the phrase is not well-formed based on occurrences of the phrase in documents that are generated by a source with the language of the phrase as the primary language of the creator.

Abstract: A neural network scheme is described that uses unsupervised learning in oscillator neural networks. Training occurs by varying the weights in proportion to the output from a frequency detector. Inputs and initial weights are split into plurality of inputs and plurality of weights. These split inputs and weights can be analog or digital. Oscillators generate signals having frequencies that represent difference in inputs, initial weights, and adjusted factors. Frequency detectors are used to compare the oscillator frequencies with a synchronized frequency of all oscillators. The output of the frequency detectors are used to generate the adjusted factors, and in turn generate trained weights.

Abstract: A method, device and a computer-readable storage medium for optimizing simulation data are provided. The method for optimizing simulation data can include: inputting the simulation data generated by a simulator to a first generative adversarial network comprising a migration model; and optimizing the simulation data generated by the simulator with the migration model to generate optimized simulation data. In an embodiment of the present application, the simulation data is optimized by the generative adversarial network to enable the simulation data closer to the real data in representation. Therefore, the quality and accuracy of the simulation data can be ensured, the validity and reliability of the simulation data can be improved at least to some extent, and the cost for constructing the simulator can also be reduced.

Abstract: A cognitive modeling system uses a cognitive model to efficiently execute a variety of tasks over large datasets. The cognitive modeling system receives an input dataset and a query specifying a task to execute in relation to the input dataset. The cognitive modeling system determines an amount of similarity between each child node of the cognitive model and one or more of the input dataset and the query, selects a particular child node with the most determined amount of similarity, and executes the task using the particular child node. The task execution includes searching the particular child node for a connected set of neurons that match a particular part of the input dataset by a threshold amount, and applying an output that is associated with the connected set of neurons to the particular part of the input dataset.

Abstract: A method for optimizing the run parameters of a software application on an information processing platform, consisting of iteratively optimizing said parameters on each execution of said application, wherein, for each execution of said application (11), a set of said parameters is determined and a run time of said application with said parameters is determined, and an association between said set and said run time is stored in order to create a history (25); and wherein said set is determined by implementing a genetic optimization algorithm comprising a step (21) consisting of selecting two sets of parameters from said history; a step (22) consisting of creating a new set of parameters by recombining said two sets of parameters; and a step (23) of random mutation of said new set of parameters.

Abstract: Methods and systems for determining an optimized setting for one or more process parameters of a machine learning training process. One of the methods includes processing a current network input using a recurrent neural network in accordance with first values of the network parameters to obtain a current network output, obtaining a measure of the performance of the machine learning training process with an updated setting defined by the current network output, and generating a new network input that comprises (i) the updated setting defined by the current network output and (ii) the measure of the performance of the training process with the updated setting defined by the current network output.

Abstract: Provided herein is a method for determining a soil classification comprising: obtaining a soil sample; conducting two or more field tests on the soil sample to obtain raw data for each field test; and calculating the soil classification from the raw data by applying a previously obtained validation dataset obtained from a training and validation soil classification calculation using samples of known soil classification, wherein the validation dataset is obtained using a feed-forward backpropagation neural network.

Abstract: Methods and systems are provided for neural architecture search. A computer-implemented neural architecture search may be used for providing a neural network configured to perform a selected task. A computational graph is obtained, which includes a plurality of nodes, edges and weightings associated with the nodes and/or edges. The computational graph includes a plurality of candidate models in the form of subgraphs of the computational graph. Selected subgraphs may be trained sequentially, with the weightings corresponding to each said subgraph being updated in response to training. For each weighting in a subgraph which is shared with another subgraph, updates to the weightings are controlled based on an indication of how important to another subgraph a node/edge associated with that weighting is.

Abstract: A method, system, and computer program product is provided for embedding compression and reconstruction. The method includes receiving embedding vector data comprising a plurality of embedding vectors. A beta-variational autoencoder is trained based on the embedding vector data and a loss equation. The method includes determining a respective entropy of a respective mean and a respective variance of each respective dimension of a plurality of dimensions. A first subset of the plurality of dimensions is determined based on the respective entropy of the respective mean and the respective variance for each respective dimension of the plurality of dimensions. A second subset of the plurality of dimensions is discarded based on the respective entropy of the respective mean and the respective variance for each respective dimension of the plurality of dimensions. The method includes generating a compressed representation of the embedding vector data based on the first subset of dimensions.

Abstract: The present invention makes it possible to reduce the burden on a user in vehicle testing while maintaining precision of vehicle testing, and is a vehicle testing system for testing, on a test bench, a vehicle or a test piece that is a portion of a vehicle, the vehicle testing system being provided with a camera for capturing an image of a portion of a test piece, and a control device for controlling vehicle testing on the basis of the image captured by the camera.

Abstract: Various embodiments of the present invention provide methods, apparatus, systems, computing devices, computing entities, and/or the like for performing natural language processing operations using an attention-based text encoder machine learning model that is trained using a multi-task training routine that is associated with two or more training tasks (e.g., a multi-task training routine that is associated with two or more sequential training tasks, a multi-training routine that is associated with two or more concurrent training tasks, and/or the like).

Abstract: Embodiments relate to a neural processor circuit with scalable architecture for instantiating one or more neural networks. The neural processor circuit includes a data buffer coupled to a memory external to the neural processor circuit, and a plurality of neural engine circuits. To execute tasks that instantiate the neural networks, each neural engine circuit generates output data using input data and kernel coefficients. A neural processor circuit may include multiple neural engine circuits that are selectively activated or deactivated according to configuration data of the tasks. Furthermore, an electronic device may include multiple neural processor circuits that are selectively activated or deactivated to execute the tasks.

Abstract: Systems, appliances, and methods for operating a cooking result inference system are provided herein. The cooking engagement system can include a controller in operable communication with a camera assembly, a mass sensor assembly, and a thermal sensor assembly. An image signal including an image of an object can be accessed. A mass signal including a mass of the object can be accessed from the mass sensor assembly. A temperature signal including a temperature of the object can be accessed from the thermal sensor assembly. An inferred cooking result can then be generated based on a machine-learned model that can perform operations on input including the image signal, the mass signal, and the temperature signal. The inferred cooking result can include an inferred image depicting the object as it is predicted to appear after a recommended cooking time at a recommended temperature.

Abstract: Methods, apparatus, and systems are disclosed for estimating panel attrition. An example apparatus includes at least one memory, machine readable instructions, and processor circuitry to execute the machine readable instructions to determine a beta distribution of a non-parametric survival curve estimate based on panel meter data associated with a cohort of panelists, determine confidence intervals for a set of beta distribution parameters associated with the survival curve estimate, and output a panelist attrition estimate generated based on the confidence intervals for the survival curve estimate, the panelist attrition estimate to represent panelist retention over time based on an installation date of the panel meter.

Abstract: The present disclosure provides a data pre-processing method and device and related computer device and storage medium. By storing the target output data corresponding to the target operation into the first memory close to the processor and reducing the time of reading the target output data, the occupation time of I/O read operations during the operation process can be reduced, and the speed and efficiency of the processor can be improved.

Abstract: An image processing method comprises obtaining an input image; converting the input image or a feature map of the input image into a plurality of target input images or target feature maps, wherein a resolution of each of the target input images or the target feature maps is smaller than a resolution of the feature map of the input image or the input image, and pixels at the same position in each of the target input images or the target feature maps are of a neighborhood relationship with the input image or the feature map of the input image; processing at least a part of the plurality of target input images or target feature maps by one or more convolution blocks in a convolutional neural network; and increasing a resolution of a feature map output from the one or more convolution blocks in the convolutional neural network.

Abstract: Methods, apparatus, systems and articles of manufacture to identify a side-channel attack are disclosed. Example instructions cause one or more processors to generate an event vector based on one or more counts corresponding to tasks performed by a central processing unit; determine distances between the event vector and weight vectors of neurons in a self-organizing map; select a neuron of the neurons that results based on a determined distance; identify neurons that neighbor the selected neuron; and update at least one of a weight vector of the selected neuron or weight vectors of the neighboring neurons based on the determined distance of the selected neuron.

Abstract: The technology disclosed relates to artificial intelligence-based determination of analyte data for base calling. In particular, the technology disclosed uses input image data that is derived from a sequence of images. Each image in the sequence of images represents an imaged region and depicts intensity emissions indicative of one or more analytes and a surrounding background of the intensity emissions at a respective one of a plurality of sequencing cycles of a sequencing run. The input image data comprises image patches extracted from each image in the sequence of images. The input image data is processed through a neural network to generate an alternative representation of the input image data. The alternative representation is processed through an output layer to generate an output indicating properties of respective portions of the imaged region.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for predicted brain data of a patient. One of the methods includes receiving montage configuration data for a specified montage; receiving raw EEG data captured using the specified montage from a brain of a particular subject; generating, using the montage configuration data and the raw EEG data, EEG connectivity data for the specified montage; using a generative neural network to map the EEG connectivity data to predicted fMRI connectivity data, the generative neural network having been trained using training EEG-fMRI connectivity data pairs, each pair comprising EEG connectivity data of a subject and fMRI connectivity data of the same subject; and taking an action based on the predicted fMRI connectivity data.

Abstract: Systems and methods include a computer-implemented method for random selection and use of observation cells. Observation cells are randomly selected from a model of process-based reactive transport modeling (RTM). The observation cells are incorporated into a neural network for proxy modeling. A set of parameter-specific proxy models represented by a neural network is trained. Each parameter-specific proxy model corresponds to a specific RTM parameter from a set of RTM parameters. Blind tests are performed using the set of parameter-specific proxy models, where each blind test tests a specific one of the parameter-specific proxy models. Predictions are generated using the set of parameter-specific proxy models. 3-dimensional interpolation the observation cells is performed.

Abstract: Systems and methods for evaluating a subsurface region of the earth for hydrocarbon exploration, development, or production are disclosed. Embodiments of the present disclosure are configured to determine advanced radioactive formation data from commonly acquired well logging data sets. In particular, a predictive model is trained to generate “synthetic” spectral gamma ray logs are from basic neutron, density and total gamma ray logs measured from a well within the formation. The predictive model comprises a neural network that is trained using multi-resolution graph clustering techniques to correlate patterns in the density, neutron and gamma ray log data to patterns in spectral gamma ray log data. Embodiments of the present disclosure are further configured to use the synthetic spectral gamma ray logs output by the model to quantify the clay content of the formation, its permeability and determine a hydrocarbon productivity index for the formation.

Abstract: Disclosed are a method for implementing and developing a network model and a related product. The method includes: an establishment command of a neural network model is received, and a computation graph of an initial neural network model is established accordingly; a computing module selected and a connection relationship of the computing module are acquired, the computing module and the connection relationship are added into the computation graph of the initial neural network model, and an intermediate neural network model is obtained; and an establishment ending command of the neural network model is collected, the intermediate neural network model is verified according to the ending command to determine whether the computing module conflicts with other computing modules or not; if not, an established neural network model matched with a computation graph of the intermediate network model is generated, and execution codes matched with the computation graph are generated.

Abstract: Systems and methods may generate a boundary of a FOU for an interval type-2 MF based on a transformation of another boundary of the FOU. The systems and methods may receive a plurality of parameters for a type-1 MF that defines a boundary of the FOU for the interval type-2 MF and may receive at least one other parameter. The systems and methods may generate, based on a transformation of the type-1 MF utilizing the at least one parameter, a type-1 MF that defines a different boundary of the FOU. The system and methods may adjust the plurality of parameters and the at least one second parameter to adjust the FOU for use in a model representing, for example, a real-world physical system, where execution of the model executes a fuzzy inference system and generates results representing a behavior of the real-world physical system.

Abstract: According to a method and apparatus for neural network quantization, a quantized neural network is generated by performing learning of a neural network, obtaining weight differences between an initial weight and an updated weight determined by the learning of each cycle for each of layers in the first neural network, analyzing a statistic of the weight differences for each of the layers, determining one or more layers, from among the layers, to be quantized with a lower-bit precision based on the analyzed statistic, and generating a second neural network by quantizing the determined one or more layers with the lower-bit precision.

Abstract: A method includes obtaining a plurality of master sensor responses with a master sensor in a set of training fluids and obtaining node sensor responses in the set of training fluids. A linear correlation between a compensated master data set and a node data set is then found for a set of training fluids and generating node sensor responses in a tool parameter space from the compensated master data set on a set of application fluids. A reverse transformation is obtained based on the node sensor responses in a complete set of calibration fluids. The reverse transformation converts each node sensor response from a tool parameter space to the synthetic parameter space and uses transformed data as inputs of various fluid predictive models to obtain fluid characteristics. The method includes modifying operation parameters of a drilling or a well testing and sampling system according to the fluid characteristics.

Abstract: A system and method for algorithmically orchestrating conversational dialogue transitions within an automated conversational system may include extracting a set of slots defined within a plurality of utterances and converting the plurality of utterances to a plurality of skeleton utterances. The method may also include grouping the plurality of skeleton utterances into a plurality of skeleton utterance groups, identifying a plurality of valid slot transition pairs based on an assessment of the plurality of skeleton utterance groups, and deriving a plurality of slot ontology groups based on the plurality of valid slot transition pairs. The system and method may use the plurality of distinct slot ontology groups and the plurality of skeleton utterances to facilitate contextually relevant dialogue transitions in the automated conversational system.

Abstract: In one respect, there is provided a system for training a neural network adapted for classifying one or more scripts. The system may include at least one processor and at least one memory. The memory may include program code which when executed by the at least one memory provides operations including: receiving a disassembled binary file that includes a plurality of instructions; processing the disassembled binary file with a convolutional neural network configured to detect a presence of one or more sequences of instructions amongst the plurality of instructions and determine a classification for the disassembled binary file based at least in part on the presence of the one or more sequences of instructions; and providing, as an output, the classification of the disassembled binary file. Related computer-implemented methods are also disclosed.

Abstract: The present disclosure discloses a system and method for distributed neural network training. The method includes: computing, by a plurality of heterogeneous computation units (HCUs) in a neural network processing system, a first plurality of gradients from a first plurality of samples; aggregating the first plurality of gradients to generate an aggregated gradient; computing, by the plurality of HCUs, a second plurality of gradients from a second plurality of samples; aggregating, at each of the plurality of HCUs, the aggregated gradient with a corresponding gradient of the second plurality of gradients to generate a local gradient update; and updating, at each of the plurality of HCUs, a local copy of a neural network with the local gradient update.

Abstract: A machine-learning accelerator system, comprising: a plurality of controllers each configured to traverse a feature map with n-dimensions according to instructions that specify, for each of the n-dimensions, a respective traversal size, wherein each controller comprises: a counter stack comprising counters each associated with a respective dimension of the n-dimensions of the feature map, wherein each counter is configured to increment a respective count from a respective initial value to the respective traversal size associated with the respective dimension associated with that counter; a plurality of address generators each configured to use the respective counts of the counters to generate at least one memory address at which a portion of the feature map is stored; and a dependency controller computing module configured to (1) track conditional statuses for incrementing the counters and (2) allow or disallow each of the counters to increment based on the conditional statuses.

Abstract: System and method for synthesizing a controller for a dynamical system includes a feeder neural network trained to estimate an ordinary differential equation (ODE) from time series training data (X) of a trajectory having embedded angular data and configured to learn dynamics of a physical system by encoding a generalization of a Hamiltonian representation of the dynamics using a constant external control term (u). A neural ODE solver receives the estimate of the ODE from the feeder neural network and synthesizes a controller to control the system to track a reference configuration.

Abstract: A system and method are disclosed including a computer and a processor and memory. The computer receives historical sales data comprising aggregated sales data for one or more items from one or more store for at least one past time period. The computer further trains a cyclic boosting model to learn model parameters by iteratively calculating for each feature and each bin factors for at least one full feature cycle. The computer further predicts one or more demand quantities during a prediction period by applying a prediction model to historical supply chain data, wherein a training period is earlier than the prediction period, and each of the one or more demand quantities is associated with at least one item of the one or more items and at least one stocking location of the one or more stocking locations during the prediction period and rendering a demand prediction feature explanation visualization.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for processing cell images using neural networks. One of the methods includes obtaining data comprising an input image of one or more biological cells illuminated with an optical microscopy technique; processing the data using a stained cell neural network; and processing the one or more stained cell images using a cell characteristic neural network, wherein the cell characteristic neural network has been configured through training to receive the one or more stained cell images and to process the one or more stained cell images to generate a cell characteristic output that characterizes features of the biological cells that are stained in the one or more stained cell images.

Abstract: A convolution calculation apparatus applied for convolution calculation of a convolution layer includes a decompression circuit, a data combination circuit and a calculation circuit. The decompression circuit decompresses compressed weighting data of a convolution kernel of the convolution layer to generate decompressed weighting data. The data combination circuit combines the decompressed weighting data and non-compressed data of the convolution kernel to restore a data order of weighting data of the convolution kernel. The calculation circuit performs calculation according to the weighting data of the convolution kernel and input data of the convolution layer.

Abstract: According to some disclosed embodiments an action is performed by an electronic social agent. The electronic social agent collects a first dataset indicating the user's state, the user's environment state, and a first user response to the performed action. Then, it is determined whether it is desirable to collect a second response from the user and, if so, it is further determined whether to generate a question to be presented to the user based on an analysis of a first dataset and the first user response. Then, an optimal time for presenting the question to the user is determined. A question that is based on the collected data and the first user response is generated by the electronic social agent for actively collecting an additional user response. Then, based on the collected additional user response, the decision-making model of the electronic social agent is updated and improved.

Abstract: Given an input image, an image enhancement task, and no external examples available to train on, an Image-Specific Deep Network is constructed tailored to solve the task for this specific image. Since there are no external examples available to train on, the network is trained on examples extracted directly from the input image itself. The current solution solves the problem of Super-Resolution (SR), whereas the framework is more general and is not restricted to SR.

Abstract: A sensor includes: a correlation matrix calculator which calculates a correlation matrix from biological information extracted using reception signals obtained by receiving, in a predetermined period, signals transmitted to a predetermined space; a first number information calculator which calculates first number information; a MUSIC spectrum calculator which estimates a candidate of a position of at least one living body and outputs a likelihood spectrum; and a second number information calculator which estimates a position or second number information, which is a total number of living bodies with an increased degree of likelihood, from first position information possibly including a plurality of position candidates.

Abstract: The present invention discloses a path planning method, including the following steps: establishing an empirical map and a corresponding episodic cognitive map using a RatSLAM algorithm based on an episodic memory model; extracting a road edge in a historical memory image with a Canny operator; performing conversion to a world coordinate system from a pixel coordinate system based on the road edge, and preliminarily judging connectivity according to slope of the road edge; continuously injecting energy into the potential path detection network according to continuous observation of a potential path, so as to further judge the road connectivity; fusing the detected potential path and the original episodic cognitive map, and correspondingly updating the empirical map; and planning a path based on the updated episodic cognitive map. The potential safe path in an environment may be detected, and a better path may be planned based on the updated episodic memory model.

Abstract: There is included a method and apparatus comprising computer code configured to cause a processor or processors to perform obtaining an input low resolution (LR) image comprising a height, a width, and a number of channels, implementing a feature learning deep neural network (DNN) configured to compute a feature tensor based on the input LR image, generating, by an upscaling DNN, a high resolution (HR) image, having a higher resolution than the input LR image, based on the feature tensor computed by the feature learning DNN, wherein a networking structure of the upscaling DNN differs depending on different scale factors, and wherein a networking structure of the feature learning DNN is a same structure for each of the different scale factors.

Abstract: A method, a computer program with instructions, and a device for predicting a course of a road based on radar data of a motor vehicle. The radar data to be processed is received and then accumulated in a measuring grid. Subsequently, clusters are formed for objects in the measuring grid. Cluster descriptions are generated for the clusters. The resulting clusters are processed to determine polynomials for describing the road edges. The polynomials are finally output for further use.

Abstract: To provide a robust information processing system against attacks by Adversarial Example. A neural network model 608, a latent space database 609 for storing position information in a latent space in which first output vectors, which are output vectors of a predetermined hidden layer included in the neural network model, are embedded concerning input data used for learning of the neural network model, and an inference control unit 606 for making an inference using the neural network model and the latent space database are provided. The inference control unit infers the input data based on the positional relationship between the second output vector, which is an output vector of the predetermined hidden layer concerning input data to be inferred, and the first output vectors in said latent space.