Abstract: Provided an apparatus that receives time series data from a data storage unit storing time series of sample data or feature values calculated from the sample data, computes a measure indicating change and repetition characteristics of the time series data, based on sample value distribution thereof, selects a state model structure to be used for model learning and estimation, from state models including a fully connected state model and a one way direction state model, based on the measure and stores the selected state model in a model storage unit.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for selecting actions to be performed by an agent interacting with an environment. In one aspect, a method comprises: obtaining a graph of nodes and edges that represents an interaction history of the agent with the environment; generating an encoded representation of the graph representing the interaction history of the agent with the environment; processing an input based on the encoded representation of the graph using an action selection neural network, in accordance with current values of action selection neural network parameters, to generate an action selection output; and selecting an action from a plurality of possible actions to be performed by the agent using the action selection output generated by the action selection neural network.

Abstract: A computer-implemented method, a computer program product, and a computer system for multi-sample dropout in deep neural network training. A computer creates multiple dropout samples in a minibatch, starting from a dropout layer and ending at a loss function layer in a deep neural network. At the dropout layer in the deep neural network, the computer applies multiple random masks for respective ones of the multiple dropout samples. At a fully connected layer in the deep neural network, the computer applies a shared parameter for all of the multiple dropout samples. After the loss function layer in the deep neural network, the computer calculates a final loss value, by averaging loss values of the respective ones of the multiple dropout samples.

Abstract: A system and method for autonomously defining regions of interest for a container are provided. The system comprises a platform for supporting the container, a detector for capturing feature data of the container while on the platform, and a computer system. The computer system is in communication with the detector and platform. The computer system is programmed to locate features of the container from the captured feature data, and define the regions of interest for the container based on the located features.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for training a student neural network. In one aspect, there is provided a method comprising: processing a training input using the student neural network to generate an output for the training input; processing the student neural network output using a discriminative neural network to generate a discriminative score for the student neural network output, wherein the discriminative score characterizes a prediction for whether the network input was generated using: (i) the student neural network, or (ii) a brain emulation neural network; and adjusting current values of the student neural network parameters using gradients of an objective function that depends on the discriminative score for the student neural network output.

Abstract: A system and method for inserting a composited image or otherwise generated graphic into a selected video by way of a programmatic process. According to some embodiments, a system may comprise an Automated Placement Opportunity Identification (APOI) engine, a Placement Insertion Interface (PII) engine, a preview system, and an automated compositing service. The system finalizes a graphic composite into a video and provides a user with a preview for final export or further manipulation.

Abstract: The present disclosure discloses a system and a method that includes receiving, at a decoder, a latent representation of an image having a first domain, and generating a reconstructed image having a second domain, wherein the reconstructed image is generated based on the latent representation.

Abstract: Systems and methods include processors for receiving training data for a user activity; receiving bias criteria; determining a set of model parameters for a machine learning model including: (1) applying the machine learning model to the training data; (2) generating model prediction errors; (3) generating a data selection vector to identify non-outlier target variables based on the model prediction errors; (4) utilizing the data selection vector to generate a non-outlier data set; (5) determining updated model parameters based on the non-outlier data set; and (6) repeating steps (1)-(5) until a censoring performance termination criterion is satisfied; training classifier model parameters for an outlier classifier machine learning model; applying the outlier classifier machine learning model to activity-related data to determine non-outlier activity-related data; and applying the machine learning model to the non-outlier activity-related data to predict future activity-related attributes for the user activity

Abstract: Systems and methods for integrated multi-factor multi-label analysis include using one or more deep learning systems, such as neural networks, to analyze how well one or more entities are likely to benefit from a targeted action. Data associated with each of the entities is analyzed to determine a score for each of the proposed targeted actions using multiple analysis factors. The scores for each analysis factor are determined using a different multi-layer analysis network for each analysis factor. The scores for each analysis factor are then combined to determine an overall score for each of the proposed targeted actions. The entities and the proposed targeted actions with the highest scores are then identified and then used to determine which entities are to be the subject of which targeted actions.

Abstract: A method provides learning with noisy labels. The method includes generating a first network of a machine learning model with a first set of parameter initial values, and generating a second network of the machine learning model with a second set of parameter initial values. First clean probabilities for samples in a training dataset are generated using the second network. A first labeled dataset and a first unlabeled dataset are generated from the training dataset based on the first clean probabilities. The first network is trained based on the first labeled dataset and first unlabeled dataset to update parameters of the first network.

Abstract: Provided is a method for implementing reinforcement learning by a neural network. The method may include performing, for each epoch of a first predetermined number of epochs, a second predetermined number of training iterations and a third predetermined number of testing iterations using a first neural network. The first neural network may include a first set of parameters, the training iterations may include a first set of hyperparameters, and the testing iterations may include a second set of hyperparameters. The testing iterations may be divided into segments, and each segment may include a fourth predetermined number of testing iterations. A first pattern may be determined based on at least one of the segments. At least one of the first set of hyperparameters or the second set of hyperparameters may be adjusted based on the pattern. A system and computer program product are also disclosed.

Abstract: A map fusing method includes receiving a source graph and a target graph. The source graph is representative of a source map and the target graph is representative of a target map and includes nodes and edges that connect the nodes. The method further includes processing each of the source graph and the target graph in a graph convolutional layer to provide graph convolutional layer outputs related to the source graph and to the target graph, processing each of the graph convolutional layer outputs for the source graph and the target graph in a linear rectifying layer to output node feature maps related to the source graph and the target graph. The method further includes selecting pairs of node representations from the node feature maps related to the source graph and the target graph and concatenating the selected pairs to output selected and concatenated pairs of node representations.

Abstract: The disclosure relates to a system and method of configuring and validating multi-vendor and multi-region Internet-of-Things (IoT) devices using reinforcement learning. In some embodiments, the method includes generating a matching table for each of a plurality of IoT sensors based on a plurality of sensor attributes extracted from a product data associated with an IoT sensor; acquiring an identification information and operational information associated with the IoT sensor and a set of neighboring IoT sensors for each of the plurality of IoT sensors; identifying an appropriate set of IoT sensors from the plurality of IoT sensors, based on a user requirement, the matching table, the identification information and the operational information, using a Reinforcement Learning (RL) model; and dynamically configuring each of the appropriate set of IoT sensors based on a vendor type.

Abstract: A portable electronic device is provided. The electronic device includes a transparent front glass cover including a planar portion that forms a front surface of the electronic device, a planar rear glass cover that forms a rear surface of the electronic device, a metal bezel that surrounds a space formed by the front glass cover and the rear glass cover, and a flexible display device that is embedded in the space and exposed through the front glass cover. The front cover includes a left bent portion and a right bent portion on the left and right of the planar portion at the center of the front cover.

Abstract: An apparatus including circuitry configured to determine a probability by combining at least: a probability that an event is present within a current feature of interest given a first set of previous features of interest, and a probability that the event is present within the current feature of interest given a second set of previous features of interest, different to the first set of previous features of interest; circuitry configured to detect the event based on the determined probability; and circuitry configured to control, in dependence on the detection of the event, performance of an action.

Abstract: Aspects of the present disclosure relate to processing irregularly arranged characters. An image is received. An irregularly arranged character within the image is detected. A direction of the irregularly arranged character is modified to a proper direction to obtain a properly oriented character. The properly oriented character is recognized to obtain a first identified character. The image is then rebuilt by replacing the irregularly arranged character with the first identified character, the first identified character in a machine-encoded format.

Abstract: Techniques are described for providing location-based information and functionality to people and computing devices in various ways. In at least some situations, the techniques include enabling multiple people in a common geographic area to interact in various ways, such as via devices capable of communications (e.g., cellular telephones, computing devices with wired and/or wireless communications capabilities, etc.), while in other situations at least some users who are remote from a particular geographic area may be allowed to intercommunicate with one or more other users or other entities in or related to that geographic area. In addition, the techniques include enabling the creation and maintenance of location-based virtual groups of users (also referred to as “clouds”), such as for users of mobile and/or fixed-location devices. Such clouds may enable various types of interactions between group members, and may be temporary and/or mobile.

Abstract: Techniques for providing audio services to multiple devices are described. For instance, connections between a hands-free unit and multiple wireless devices are established. The connections are themselves used to establish active communication channels, such as active audio communication channels, between the hands-free unit and the wireless devices, such as during a phone call. Upon establishment of an active communication channel with one of the wireless devices, the connections to the other wireless devices are disconnected—and/or additional connections refused—for the duration of the active communication channel. Furthermore, a routing module in various embodiments permits multiple hands-free units to route active communication channels to each other depending on user location.

Abstract: A computer system trains a neural network on an instance segmentation task by casting the problem as one of mapping each pixel to a probability distribution over arbitrary instance labels. This simplifies both the training and inference problems, because the formulation is end-to-end trainable and requires no post-processing to extract maximum a posteriori estimates of the instance labels.

Abstract: Systems and methods are described in relation to specific technical improvements adapted for machine learning architectures that conduct classification on numerical and/or unstructured data. In an embodiment, two neural networks are utilized in concert to generate output data sets representative of predicted future states of an entity. A second learning architecture is trained to cluster prior entities based on characteristics converted into the form of features and event occurrence such that a boundary function can be established between the clusters to form a decision boundary between decision regions. These outputs are mapped to a space defined by the boundary function, such that the mapping can be used to determine whether a future state event is likely to occur at a particular time in the future.