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: Techniques for forecasting using deep factor models with random effects are described. A forecasting framework combines the strengths of both classical and neural forecasting methods in a global-local framework for forecasting multiple time series. A global model captures the common latent patterns shared by all time series, while a local model explains the variations at the individual level.

Abstract: A computer-implemented method, computer program product, and computer processing system are provided for obtaining a plurality of bad demonstrations. The method includes reading, by a processor device, a protagonist environment. The method further includes training, by the processor device, a plurality of antagonist agents to fail a task by reinforcement learning using the protagonist environment. The method also includes collecting, by the processor device, the plurality of bad demonstrations by playing the trained antagonist agents on the protagonist environment.

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: 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: 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: Machine learning models, semantic networks, adaptive systems, artificial neural networks, convolutional neural networks, and other forms of knowledge processing systems are disclosed. Input data for a machine learning system may be analyzed to determine one or more potential biases in the input data. Based on the one or more potential biases, the input data may be grouped, and/or weights may be applied to one or more portions of the input data. The input data may be input into a machine learning algorithm, which may generate output data. Based on an evaluation of the output data, the input data may be grouped, and/or second weights may be applied to one or more portions of the input data.

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: 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: A method of communicating information, comprising modeling a stream of sensor data, to produce parameters of a predictive statistical model; communicating information defining the predictive statistical model from a transmitter to a receiver; and after communicating the information defining the predictive statistical model to the receiver, communicating information characterizing subsequent sensor data from the transmitter to the receiver, dependent on an error of the subsequent sensor data with respect to a prediction of the subsequent sensor data by the statistical model. A corresponding method is also encompassed.

Abstract: The present disclosure relates to systems, non-transitory computer-readable media, and methods for generating user embeddings utilizing an interaction-to-vector neural network. For example, a user embeddings system transforms unorganized data of user interactions with content items into structured user interaction data. Further, the user embeddings system can utilize the structured user interaction data to train a neural network in a semi-supervised manner and generate uniform vectorized user embeddings for each of the users.

Abstract: A method and apparatus of open set recognition, and a computer-readable storage medium are disclosed. The method comprises acquiring auxiliary data and training data of known categories for open set recognition, training a neural network alternately using the auxiliary data and the training data, until convergence; extracting a feature of data to be recognized for open set recognition, using the trained neural network; and recognizing a category of data to be recognized, based on the feature of the data to be recognized.

Abstract: Described is a system (and method) for training and using machine learning models to identify potential discrepancies between predicted odds and actual odds for a future event. The system may create a predicted odds machine learning model using an ensemble training algorithm. To create the risk management machine learning model, the system may determine a set of past odds differences between predicted odds outputted by the predicted odds machine learning model for past events and the actual historical odds for those past events. Once the predicted odds machine learning model and risk management machine learning model are trained, the system may use current (or real-time) event information to determine potential opportunities to leverage based on discrepancies between predicted odds of upcoming events and the actual odds for those events.

Abstract: A method for natural language processing includes receiving, by one or more processors, an unstructured text input. An entity classifier is used to identify entities in the unstructured text input. The identifying the entities includes generating, using a plurality of sub-classifiers of a hierarchical neural network classifier of the entity classifier, a plurality of lower-level entity identifications associated with the unstructured text input. The identifying the entities further includes generating, using a combiner of the hierarchical neural network classifier, a plurality of higher-level entity identifications associated with the unstructured text input based on the plurality of lower-level entity identifications. Identified entities are provided based on the plurality of higher-level entity identifications.

Abstract: A chromosome recognition method based on deep learning includes the following steps: step 1, obtaining an independent chromosome image; step 2, calculating a manual feature of a chromosome; step 3, performing basic image processing on the chromosome; step 4, building a deep learning model; and step 5, predicting a type of the chromosome based on the deep learning model. By adopting a deep learning method, the chromosome recognition method can be used for recognizing the chromosome type accurately and efficiently. Compared with an existing recognition technology, the chromosome recognition method based on deep learning of the present invention has the advantages that the chromosome karyotype analysis efficiency can be effectively improved, the recognition sequencing time can be shortened, automatic classification and sequencing of chromosomes can be completely with high accuracy.

Abstract: Anomaly detection from time series is one of the key components in automated monitoring of one or more entities. Domain-driven sensor selection for anomaly detection is restricted by knowledge of important sensors to capture only a certain set of anomalies from the entire set of possible anomalies. Hence, existing anomaly detection approaches are not very effective for multi-dimensional time series. Embodiments of the present disclosure depict sparse neural network for anomaly detection in multi-dimensional time series (MDTS) corresponding to a plurality of parameters of entities. A reduced-dimensional time series is obtained from the MDTS via an at least one feedforward layer by using a dimensionality reduction model. The dimensionality reduction model and recurrent neural network (RNN) encoder-decoder model are simultaneously learned to obtain a multi-layered sparse neural network. A plurality of error vectors corresponding to at least one time instance of the MDTS is computed to obtain an anomaly score.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for an artificial intelligence system. In one aspect, a system includes multiple artificial intelligence skill agents that have each been trained to perform different actions in a telecommunications system. The system also includes an orchestrator agent that interacts with each of the artificial intelligence skill agents and coordinates which of the artificial intelligence agents performs actions in response to user inputs. The orchestrator agent receives a user input and determines an intent expressed by the user input. The orchestrator agent transmits an instruction to an artificial intelligence skill agent to perform an action that provides a response to the intent. In response to receiving the instruction from the orchestrator agent, the artificial intelligence skill agent performs the action when the artificial intelligence skill agent is capable of carrying out the action.

Abstract: The technology disclosed relates to constructing a convolutional neural network-based classifier for variant classification. In particular, it relates to training a convolutional neural network-based classifier on training data using a backpropagation-based gradient update technique that progressively match outputs of the convolutional network network-based classifier with corresponding ground truth labels. The convolutional neural network-based classifier comprises groups of residual blocks, each group of residual blocks is parameterized by a number of convolution filters in the residual blocks, a convolution window size of the residual blocks, and an atrous convolution rate of the residual blocks, the size of convolution window varies between groups of residual blocks, the atrous convolution rate varies between groups of residual blocks. The training data includes benign training examples and pathogenic training examples of translated sequence pairs generated from benign variants and pathogenic variants.