Abstract: According to an exemplary embodiment of the present disclosure, a computer program stored in a computer readable storage medium is disclosed. The computer program performs operations for processing input data when the computer program is executed by one or more processors of a computer device, the operations including: obtaining input data based on sensor data obtained during manufacturing of an article by using one or more manufacturing recipes in one or more manufacturing equipment; inputting the input data to a neural network model loaded to the computer device; generating an output by processing the input data by using the neural network model; and detecting an anomaly for the input data based on the output of the neural network model.

Abstract: Classes are identified in a dataset, and an independent artificial neural network is created for each class in the dataset. Thereafter, all classes in the dataset are provided to each independent artificial neural network. Each independent artificial neural network is separately trained to respond to a single particular class in the dataset and to reject all other classes in the dataset. Output from each independent artificial neural network is provided to a combining classifier, and the combining classifier is trained to identify all classes in the dataset based on the output of all the independent artificial neural networks.

Abstract: A neural net processor provides twin processing paths trainable using different moments of the input data, one moment providing a proxy for uncertainty. Subsequent operation of the trained neural net allows monitoring of the uncertainty proxy to provide real-time assessment of neural net model-based uncertainty.

Abstract: There are provided a neural network weight discretizing method, system and device, and a computer readable storage medium. The method includes acquiring a weight value range and a number of discrete weight states, the weight value range referring to a range of discrete weight values consisting of a maximum weight value of a current time step and a minimum weight value of the current time step, and the number of discrete weight states referring to the quantity of discrete weight states. The method also includes acquiring a weight state of a previous time step and a weight increment of the current time step and acquiring a state transfer direction by using a directional function according to the weight increment of the current time step. The method also includes acquiring a weight state of the current time step according to the weight state of the previous time step, the weight increment of the current time step, the state transfer direction, the weight value range and the number of discrete weight states.

Abstract: A system includes a neural network system implemented by one or more computers. The neural network system is configured to receive an observation characterizing a current state of a real-world environment being interacted with by a robotic agent to perform a robotic task and to process the observation to generate a policy output that defines an action to be performed by the robotic agent in response to the observation. The neural network system includes: (i) a sequence of deep neural networks (DNNs), in which the sequence of DNNs includes a simulation-trained DNN that has been trained on interactions of a simulated version of the robotic agent with a simulated version of the real-world environment to perform a simulated version of the robotic task, and (ii) a first robot-trained DNN that is configured to receive the observation and to process the observation to generate the policy output.

Abstract: A method for creating a deep neural network. The deep neural network includes a plurality of layers and connections having weights, and the weights in the created deep neural network are able to assume only predefinable discrete values from a predefinable list of discrete values. The method includes: providing at least one training input variable for the deep neural network; ascertaining a variable characterizing a cost function, which includes a first variable, which characterizes a deviation of an output variable of the deep neural network ascertained as a function of the provided training input variable relative to a predefinable setpoint output variable, and the variable characterizing the cost function further including at least one penalization variable, which characterizes a deviation of a value of one of the weights from at least one of at least two of the predefinable discrete values; training the deep neural network.

Abstract: Disclosed is a light-emitting diode which includes a light-emitting epitaxial layered unit, an insulation layer, a transparent conductive layer, a protective layer, a first electrode, and a second electrode. The light-emitting epitaxial layered unit includes a first semiconductor layer, a second semiconductor layer, and a light-emitting layer sandwiched between the first and second semiconductor layers, and has a first electrode region which includes a pad area and an extension area. The insulation layer is disposed on the first semiconductor layer and at the extension area of the first electrode region. Also disclosed is a method for manufacturing the light-emitting diode.

Abstract: Real-time three-dimensional (3D) map building method and device using a 3D lidar includes representing 3D map data of a surrounding environment acquired by using a 3D lidar attached to a moving object as voxels, acquiring an eigenvalue and an eigenvector for each voxel based on all 3D points in a 3D map represented as the voxels, detecting a 3D corresponding point in the voxel corresponding to all the 3D points of 3D data newly acquired by using the 3D lidar while the moving object travels, calculating a rotation transformation and a translation transformation for minimizing an error by minimizing an inner product value between the eigenvector weighted by the eigenvalue of the voxel to which the 3D corresponding point belongs and a vector generated from a 3D corresponding point, and updating the 3D map data based on the rotation transformation and the translation transformation.

Abstract: A semiconductor device is provided. The semiconductor device includes a substrate; a buffer layer on the substrate; a channel layer on the buffer layer; a barrier layer on the channel layer; a doped compound semiconductor layer on a portion of the barrier layer; an un-doped first capping layer on the doped compound semiconductor layer; a gate structure on the un-doped first capping layer; and source/drain structures on opposite sides of the gate structure. There is a channel region in the channel layer that is adjacent to the interface between the channel layer and the barrier layer.

Abstract: An optimization method includes holding combining destination information indicating a combining destination neuron to be combined with a target neuron which is one of a plurality of neurons corresponding to a plurality of spins of an Ising model obtained by converting an optimization problem, the target neuron being different in a plurality of neuron circuits; holding a weighting coefficient indicating a strength of combining between the target neuron and the combining destination neuron, and outputting the weighting coefficient corresponding to the combining destination information; permitting an update of a value of the target neuron by using the weighting coefficient output and the value of the update target neuron, and outputting a determination result indicating whether or not the value of the target neuron is permitted to be updated; and determining the update target neuron based on the plurality of determination results respectively output and outputting the update target information.

Abstract: A computer-implemented system and method for storing data associated with an agent in a multi-dimensional environment via a memory architecture. The memory architecture is structured so that each unique position in the environment corresponds to a unique position within the memory architecture, thereby allowing the memory architecture to store features located at a particular position in the environment in a memory location specific to that location. As the agent traverses the environment, the agent compares the features at the agent's particular position to a summary of the features stored throughout the memory architecture and writes the features that correspond to the summary to the coordinates in the memory architecture that correspond to the agent's position. The system and method allows agents to learn, using a reinforcement signal, how to behave when acting in an environment that requires storing information over long time steps.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for training a classifier to detect open vehicle doors. One of the methods includes obtaining a plurality of initial training examples, each initial training example comprising (i) a sensor sample from a collection of sensor samples and (ii) data classifying the sensor sample as characterizing a vehicle that has an open door; generating a plurality of additional training examples, comprising, for each initial training example: identifying, from the collection of sensor samples, one or more additional sensor samples that were captured less than a threshold amount of time before the sensor sample in the initial training example was captured; and training the machine learning classifier on first training data that includes the initial training examples and the additional training examples to generate updated weights for the machine learning classifier.

Abstract: Examples are disclosed that relate to computing devices, head-mounted display devices, and methods for displaying holographic objects using slicing planes or volumes. In one example a computing device causes a display system to display a holographic object associated with a holographic volume, the holographic object occluding an occluded holographic object that is not displayed. Location data of at least a portion of a hand is received from a sensor. The location data of the hand is used to locate a slicing plane or a slicing volume within the holographic volume. Based on the location of the slicing plane or the slicing volume, at least a portion of the occluded holographic object is displayed via the display system.

Abstract: The present application discloses a semiconductor device and a method for fabricating the semiconductor device. The semiconductor device includes a substrate, a gate structure positioned on the substrate, and a plurality of word lines positioned apart from the gate structure, wherein a top surface of the gate structure and top surfaces of the plurality of word lines are at a same vertical level.

Abstract: A mixed reality (MR) device can allow a user to switch between input modes to allow interactions with a virtual environment via devices such as a six degrees of freedom (6DoF) handheld controller and a touchpad input device. A default input mode for interacting with virtual content may rely on the user's head pose, which may be difficult to use in selecting virtual objects that are far away in the virtual environment. Thus, the system may be configured to allow the user to use a 6DoF cursor, and a visual ray that extends from the handheld controller to the cursor, to enable precise targeting. Input via a touchpad input device (e.g., that allows three degrees of freedom movements) may also be used in conjunction with the 6DoF cursor.

Abstract: A substrate having an electronic component embedded therein includes a core substrate including first and second wiring layers disposed on different levels and one or more insulating layers disposed between the first and second wiring layers, having a cavity in which a stopper layer is disposed on a bottom surface of the cavity, and including a groove disposed around the stopper layer on the bottom surface; an electronic component disposed on the stopper layer in the cavity; an insulating material covering at least a portion of each of the core substrate and the electronic component and disposed in at least a portion of each of the cavity and the groove; and a third wiring layer disposed on the insulating material. The stopper layer protrudes on the bottom surface.

Abstract: A system for generating a nearest neighboring vertices index. The system includes a memory and one or more processors. The one or more processors receive a base figure asset and an item asset, determine nearest neighbor vertices between the base figure asset and the item asset using at least one of a k-dimensional tree algorithm and a geodesic algorithm, and generate the nearest neighboring vertices index based on the determined nearest neighbor vertices between the base figure asset and the item asset.

Abstract: A photovoltaic device including a photovoltaic cell and method of use is disclosed. The photovoltaic cell includes at least a first photovoltaic layer and a second photovoltaic layer arranged in a stack. The first photovoltaic layer has a first thickness and receives light at its top surface. A second photovoltaic layer has a second thickness and is disposed beneath the first photovoltaic layer and receives light passing through the first photovoltaic layer. The first thickness and the second thickness are selected so that a first light absorption at the first photovoltaic layer is equal to a second light absorption at the second photovoltaic layer. The photovoltaic cell is irradiated at its top surface with monochromatic light to generate a current.

Abstract: A neural network-based rendering technique increases temporal stability and image fidelity of low sample count path tracing by optimizing a distribution of samples for rendering each image in a sequence. A sample predictor neural network learns spatio-temporal sampling strategies such as placing more samples in dis-occluded regions and tracking specular highlights. Temporal feedback enables a denoiser neural network to boost the effective input sample count and increases temporal stability. The initial uniform sampling step typically present in adaptive sampling algorithms is not needed. The sample predictor and denoiser operate at interactive rates to achieve significantly improved image quality and temporal stability compared with conventional adaptive sampling techniques.

Abstract: A virtual reality display method, device, apparatus and storage medium are provided. The method includes: acquiring multimedia data to be displayed and a visible region of a viewer on a curved display surface, wherein the visible region is obtained by projecting a visible range of the viewer to the curved display surface, and is not larger than a display area of the curved display surface; determining target curvatures of at least two positions in the visible region of the viewer, wherein in the target curvatures, target curvatures of different positions are related to a distance to a center of the visible region of the viewer; adjusting, based on the target curvatures of the at least two positions in the visible region, a curvature of a corresponding position on the curved display surface; and mapping the multimedia data to be displayed to the curved display surface having the adjusted curvature.