Abstract: Implementations of the present specification provide a method and an apparatus for extending question and answer samples. According to the method, a random number is generated for each existing sample, a question is blurred for a sample whose random number belongs to sample extension random numbers, to generate an extended sample, so that an overall sample blurring extension rate can be effectively controlled. In addition, for a sample needing blurring extension, a question is extended by deleting a word with a predetermined part of speech in the corresponding question, and then an extended sample is generated based on an extended question, so that more question expression ways are compatible. As such, a question and answer model is trained by using a sample set to which extended samples are added, so that an answer can be provided to a user more effectively.

Abstract: Systems and methods are disclosed to implement a distributed matrix decomposition system using gossip. In embodiments, the matrix decomposition system employs a scalable, parallel, and decentralized approach to divide an input matrix into a grid blocks, and individually decompose the blocks into local decomposed matrices by communicating (gossiping) with a limited set of neighboring blocks. In embodiments, the decomposition may be implemented as an iterative process using Stochastic Gradient Descent, where the decomposed matrices are iteratively updated and kept in approximate agreement for neighboring blocks. The division of the input matrix allows the decomposition operation to be easily parallelized among nodes of a distributed computing system and scaled to suit the size of the input matrix. Moreover, the distributed approach eliminates the need for a central server, which in some systems may represent an operational bottle neck, a single point of failure, or a target for attacks.

Abstract: A method, system and computer program product are disclosed for predicting influence in a social network. In one embodiment, the method comprises identifying a set of users of the social network, and identifying a subset of the users as influential users based on defined criteria. A multitude of measures are identified as predictors of which ones of the set of users are the influential users. These measures are aggregated, and a composite predictor model is formed based on this aggregation. This composite predictor model is used to predict which ones of the set of users will have a specified influence in the social network in the future. In one embodiment, the specified influence is based on messages sent from the users, and for example, may be based on the number of the messages sent from each user that are re-sent by other users.

Abstract: A device that includes a first correlithm object delay device, a second correlithm object delay device, and a third correlithm object delay device. The a first correlithm object delay device is operable to receive a first correlithm object at a first time, store the first correlithm object, and output the first correlithm object at a second time later than the first time. The second correlithm object delay device is operable to receive the first correlithm object at a third time later than the second time, store the first correlithm object, and output the first correlithm object at a fourth time later than the third time. The third correlithm object delay device is operable to receive the first correlithm object at a fifth time later than the fourth time, store the first correlithm object, and output the first correlithm object at a sixth time later than the fifth time.

Abstract: A method for implementing a convolutional neural network (CNN) accelerator on a target includes utilizing one or more processing elements to perform convolution. A configuration of the CNN accelerator is modified to change filters implemented by the CNN accelerator and to change formatting of output data. The one or more processing elements are utilized to perform one of deconvolution and backpropagation convolution in response to the change in the filters and formatting of the output data.

Abstract: The present invention discloses an integrated neuro-processor comprising at least a three-dimensional memory (3D-M) array. The 3D-M array stores the synaptic weights, while the neuro-processing circuit performs neural processing. The 3-D integration between the 3D-M array and the neuro-processing circuit not only improves the computational power per die area, but also greatly increases the storage capacity per die area.

Abstract: Embodiments for detecting data anomalies by a processor. A machine learning model may be trained according to collected scores and anomaly labels of a plurality of anomaly detection operations applied to one or more data sets such that the collected scores and labels identify a degree of accuracy of estimating anomalies for each of the plurality of anomaly detection operations. An anomaly may be detected in an unstructured data set by applying the trained machine learning model on an unstructured data set.

Abstract: A vertically integrated neuro-processor comprises a plurality of neural storage-processing units (NSPU's). Each NSPU comprises at least a neuro-storage circuit and a neuro-processing circuit. The neuro-storage circuit comprises a memory array for storing at least a synaptic weight, while the neuro-processing circuit performs neural processing with the synaptic weight. The memory array and the neuro-processing circuit are vertically stacked and communicatively coupled by a plurality of inter-level connections.

Abstract: A wireless power transmission apparatus can include a coil unit including a first layer including a first wireless power transmission coil having an asymmetric shape, and a second layer including a second wireless power transmission coil having an asymmetric shape, the first and second wireless power transmission coils being connected in parallel and partially overlapping with each other; and first and second terminals configured to simultaneously apply current to the first and second wireless power transmission coils connected in parallel, in which the first wireless power transmission coil and the second wireless power transmission coil have line symmetry with respect to an imaginary line extending between the first and second terminals and between the first and second layers.

Abstract: A method, system, and computer program product for resource management are described. The method includes selecting trouble regions within the service area, generating clustered regions, and training a trouble forecast model for the trouble regions for each type of damage, the training for each trouble region using training data from every trouble region within the clustered region associated with the trouble region. The method also includes applying the trouble forecast model for each trouble region within the service area for each type of damage, determining a trouble forecast for the service area for each type of damage based on the trouble forecast for each of the trouble regions within the service area, and determining a job forecast for the service area based on the trouble forecast for the service area, wherein the managing resources is based on the job forecast for the service area.

Abstract: In some embodiments, a computerized method may comprise: determining, by an adaptive cataclysm controller operating on one or more processors, that the population of candidate solutions has converged during a run of the genetic algorithm; determining, by the adaptive cataclysm controller operating on the one or more processors, a cause for convergence of the population based, at least in part, on an analysis of the run of the genetic algorithm; selecting, by the adaptive cataclysm controller operating on the one or more processors, a first cataclysm strategy of a plurality of different cataclysm strategies based, at least in part, on one of the cause for the convergence and a history of the run of the genetic algorithm; and simulating, by the adaptive cataclysm controller operating on the one or more processors, a cataclysm based on the first cataclysm strategy.

Abstract: Determination of a user-specific likelihood of incident occurrence at a geographic location may be performed. User information, historical incident information, contextual information, and/or other information may be obtained. User information may include user demographic information, user behavior information, user social information, and/or other user related information. Historical incident information may include data relating to crime, mortality, injury, morbidity rates and may be obtained from local law enforcement, local Departments of Motor Vehicles, national security agency such as the Federal Bureau of Investigation, foreign security agency such as the Central Intelligence Agency, international criminal policy organization such as Interpol, national public health agency such as Center for Disease Control, international public health agency such as World Health Organization and/or other sources.

Abstract: In one embodiment, a method includes receiving a request to identify a word representing a target concept that is in a first relationship with a particular concept such that the first relationship is analogous to a second relationship in which a first reference concept is with a second reference concept, accessing a table of word vector relationships, looking up a particular word vector, a first reference word vector, and a second reference word vector, determining an imaginary vector such that a first vector from the first reference word vector to the second reference word vector is equal to a second vector from the particular word vector to the imaginary vector, selecting a target word vector closest to the imaginary vector, identifying a target n-gram corresponding to the target word vector, and sending a response message comprising the target n-gram.

Abstract: Artificial neuron apparatus includes first and second resistive memory cells. The first resistive memory cell is connected in first circuitry having a first input and output. The second resistive memory cell is connected in second circuitry having a second input and output. The first and second circuitry are operable in alternating read and write phases to apply a programming current to their respective memory cells on receipt of excitatory and inhibitory neuron input signals, respectively. During the write phase, resistance of the respective cells is changed in response to successive excitatory and inhibitory neuron input signals. During the read phase, a read current is applied to their respective cells to produce first and second measurement signals, respectively. An output circuit connected to the first and second outputs produces a neuron output signal at a neuron output when a difference between the first and second measurement signals traverses a threshold.

Abstract: Systems and methods for predicting feature values in a matrix are disclosed. In example embodiments, a server accesses a matrix, the matrix having multiple dimensions, one dimension of the matrix representing features, and one dimension of the matrix representing entities. The server separates the matrix into multiple submatrices along a first dimension, each submatrix including all cells in the matrix for a set of values in the first dimension. The server provides the multiple submatrices to multiple machines. The server computes, using each machine, a correlation between values in at least one second dimension of the matrix and a value for a preselected feature in the matrix, the correlation being used to predict the value for the preselected feature based on other values along the at least one second dimension. The server provides an output representing the computed correlation.

Abstract: Methods and systems are provided for implementing axiomatic control of an automorphic dynamical system. An automorphic system architecture for a dynamical system may apply ergodic techniques for axiomatically describing networked computing environments of vast dimensions comprised of a plurality of disparate information resources, processors, memory, storage, analog and digital sensors (as examples) and presenting this plurality of information sources as a singular automorphic resource through a time series. An example system may further implement dynamic axiomatic control and modification of operational parameters, without human intervention, incorporating data from a variety of sources and platforms and over multiple dimensions of context associated with those data.

Abstract: Methods, systems, and computer programs are presented for determining cluster areas within a region having higher estimates of damage caused by an earthquake as compared to damage in nearby areas. One method includes operations for identifying features associated with buildings within the region, and for training a machine learning program based on the identified features and earthquake damage data. In addition, the method includes operations for estimating, by the machine learning program, block damage caused by an earthquake, and for identifying a critical damage area (CDA) within the region. The CDA comprises a plurality of blocks geographically clustered that have a highest value of block damage. Additionally, the method includes an operation for causing presentation of the CDA within a map of the region.

Abstract: Methods and systems for detecting disparate incidents in processed data using a plurality of machine learning models. For example, the system may receive native asset data. The system may extract telemetry data from the native asset data. The system may input the first feature input into a first machine learning model, wherein the first machine learning model is trained to detect known incidents of a first type in a first set of labeled telemetry data. The system may then detect a first incident based on a first output from the first machine learning model, wherein the first incident is a first event in an asset related to the user's behavior.

Abstract: A method, system, and computer program product for resource management are described. The method includes selecting trouble regions within the service area, generating clustered regions, and training a trouble forecast model for the trouble regions for each type of damage, the training for each trouble region using training data from every trouble region within the clustered region associated with the trouble region. The method also includes applying the trouble forecast model for each trouble region within the service area for each type of damage, determining a trouble forecast for the service area for each type of damage based on the trouble forecast for each of the trouble regions within the service area, and determining a job forecast for the service area based on the trouble forecast for the service area, wherein the managing resources is based on the job forecast for the service area.

Abstract: A training method is provided. The training method includes clustering, by a processor, a plurality of items that each have an item attribute value, according to the item attribute value. The training method further includes generating, by the processor, for each item, a cluster attribute value corresponding to a cluster associated with the item. The training method also includes training, by the processor, an estimation model for estimating selection behavior of a target with respect to a choice set including two or more items, based on the cluster attribute value associated with each item included in the choice set, by using training data that includes a group of a choice set of items presented to the target and an item selected by the target from among the choice set.