Abstract: The present invention relates to a clustering method based on iterations of neural networks, which comprises the following steps: step 1, initializing parameters of an extreme learning machine; step 2, randomly choosing samples of which number is equal to the number of clusters, each sample representing one cluster, forming an initial exemplar set and training the extreme learning machine; step 3, using current extreme learning machine to cluster samples, which generates a clustering result; step 4, choosing multiple samples from each cluster as exemplars for the cluster according to a rule; step 5, retraining the extreme learning machine by using the exemplars for each cluster obtained from step 4; and step 6, going back to step 3 to do iteration, otherwise obtaining and outputting clustering result until clustering result is steady or a maximal limit of the number of iterations is reached.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for batching inputs to machine learning models. One of the methods includes receiving a stream of requests, each request identifying a respective input for processing by a first machine learning model; adding the respective input from each request to a first queue of inputs for processing by the first machine learning model; determining, at a first time, that a count of inputs in the first queue as of the first time equals or exceeds a maximum batch size and, in response: generating a first batched input from the inputs in the queue as of the first time so that a count of inputs in the first batched input equals the maximum batch size, and providing the first batched input for processing by the first machine learning model.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for stream processing. One method includes receiving an event stream of events by a first plurality of local modelers of a stream processing system. Each local modeler processes a portion of received events of the event stream according to a first set of operations, the operations including aggregating information associated with each event to generate aggregated information. One or more local modelers provide, to a first central modeler executing on the system, the respective aggregated information generated by one or more of the local modelers. A set of parameters of a respective machine learning model is determined using the received aggregated information.

Abstract: Techniques for multimodal data fusion having a multimodal hierarchical dictionary learning framework that learns latent subspaces with hierarchical overlaps are provided. In one aspect, a method for multi-view data fusion with hierarchical multi-view dictionary learning is provided which includes the steps of: extracting multi-view features from input data; defining feature groups that group together the multi-view features that are related; defining a hierarchical structure of the feature groups; and learning a dictionary using the feature groups and the hierarchy of the feature groups. A system for multi-view data fusion with hierarchical multi-view dictionary learning is also provided.

Abstract: A machine learning system builds and uses computer models for identifying how to evaluate the level of success reflected in a recorded observation of a task. Such computer models may be used to generate a policy for controlling a robotic system performing the task. The computer models can also be used to evaluate robotic task performance and provide feedback for recalibrating the robotic control policy.

Abstract: A machine learning system builds and uses computer models for identifying how to evaluate the level of success reflected in a recorded observation of a task. Such computer models may be used to generate a policy for controlling a robotic system performing the task. The computer models can also be used to evaluate robotic task performance and provide feedback for recalibrating the robotic control policy.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for keyword spotting. One of the methods includes training, by a keyword detection system, a convolutional neural network for keyword detection by providing a two-dimensional set of input values to the convolutional neural network, the input values including a first dimension in time and a second dimension in frequency, and performing convolutional multiplication on the two-dimensional set of input values for a filter using a frequency stride greater than one to generate a feature map.

Abstract: Machine-learning models and behavior can be visualized. For example, a machine-learning model can be taught using a teaching dataset. A test input can then be provided to the machine-learning model to determine a baseline confidence-score of the machine-learning model. Next, weights for elements in the teaching dataset can be determined. An analysis dataset can be generated that includes a subset of the elements that have corresponding weights above a predefined threshold. For each overlapping element in both the analysis dataset and the test input, (i) a modified version of the test input can be generated that excludes the overlapping element, and (ii) the modified version of the test input can be provided to the machine-learning model to determine an effect of the overlapping element on the baseline confidence-score. A graphical user interface can be generated that visually depicts the test input and various elements' effects on the baseline confidence-score.

Abstract: A dynamic time-evolution Boltzmann machine capable of learning is provided. Aspects include acquiring a time-series input data and supplying a plurality of input values of input data of the time-series input data at one time point to a plurality of nodes of the mode. Aspects also include computing, based on an input data sequence before the one time point in the time-series input data and a weight parameter between each of a plurality of input values of input data of the input data sequence and a corresponding one of the plurality of nodes of the model, a conditional probability of the input value at the one time point given that the input data sequence has occurred. Aspects further include adjusting the weight parameter so as to increase a conditional probability of occurrence of the input data at the one time point given that the input data sequence has occurred.

Abstract: A cleaning process optimization device includes a machine learning device that learns a cleaning condition when cleaning an object to be cleaned. The machine learning device includes: a state observation unit that observes cleaning condition data indicating the cleaning condition, and contamination state data indicating a contamination state of the object to be cleaned measured before cleaning as a state variable representing a current state of environment; a determination data acquisition unit that acquires determination data indicating an adequacy determination result on accuracy of a contamination state of the object to be cleaned after cleaning; and a learning unit that learns the cleaning condition when cleaning the object to be cleaned in association with the contamination state data using the state variable and the determination data.

Abstract: Methods, systems, and apparatus for improving exchange systems. In one aspect, a method includes receiving data representing an exchange problem; determining, from the received data, an integer programming formulation of the exchange problem; mapping the integer programming formulation of the exchange problem to a quadratic unconstrained binary optimization (QUBO) formulation of the exchange problem; obtaining data representing a solution to the exchange problem from a quantum computing resource; and initiating an action based on the obtained data representing a solution to the exchange problem.

Abstract: A system for generating a graphical user interface in a client device. The system may include a processor in communication with the client device and a database. The processor may execute: receiving a request for occupancy information of a specified merchant; obtaining a plurality of credit card authorizations associated with the merchant; generating a posted transaction array based on the credit card authorizations; removing outlier members of the posted transaction array by applying a threshold filter; generating a transaction frequency array based on the posted transaction array, the transaction frequency array comprising weekdays and aggregated transactions associated with the weekdays; modifying the transaction frequency array by applying a transformation to the aggregated transactions; generating a smoothed array by applying a kernel density estimate to the transaction frequency array; and generating a graphical user interface displaying information in the smoothed array.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for stream processing. One method includes receiving an event stream of first events by a first plurality of first local modelers of a stream processing system. Each local modeler processes a portion of received events of the event stream according to a first set of operations, the operations including aggregating information associated with each event to generate first aggregated information. A second plurality of second local modelers similarly generates second aggregated information from an event stream of second events. First and second local modelers provide, to a first central modeler, first and second aggregated information. A set of parameters of a respective machine learning model is determined by the first central modeler using the received aggregated information.

Abstract: A method for learning a classification model by a computer system is disclosed. One or more positive class data and one or more negative class data are prepared. The classification model is trained based on the positive class data to adjust one or more parameters of the classification model so that the positive class data is reconstructed by the classification model. The classification model is trained based on the negative class data to adjust the one or more parameters so that the negative class data is prevented from being reconstructed by the classification model. For the negative class data, changes in the one or more parameters with gradient of an objective function may be calculated using an unsupervised learning algorithm. The one or more parameters may be updated based on the changes in an opposite manner to the training based on the positive class.

Abstract: A learning method for a CNN (Convolutional Neural Network) capable of encoding at least one training image with multiple feeding layers, wherein the CNN includes a 1st to an n-th convolutional layers, which respectively generate a 1st to an n-th main feature maps by applying convolution operations to the training image, and a 1st to an h-th feeding layers respectively corresponding to h convolutional layers (1?h?n?1)) is provided. The learning method includes steps of: a learning device instructing the convolutional layers to generate the 1st to the n-th main feature maps, wherein the learning device instructs a k-th convolutional layer to acquire a (k?1)-th main feature map and an m-th sub feature map, and to generate a k-th main feature map by applying the convolution operations to the (k?1)-th integrated feature map generated by integrating the (k?1)-th main feature map and the m-th sub feature map.

Abstract: An intuitive object-generation experience is provided by employing an autoencoder neural network to reduce the dimensionality of a procedural model. A set of sample objects are generated using the procedural model. In embodiments, the sample objects may be selected according to visual features such that the sample objects are uniformly distributed in visual appearance. Both procedural model parameters and visual features from the sample objects are used to train an autoencoder neural network, which maps a small number of new parameters to the larger number of procedural model parameters of the original procedural model. A user interface may be provided that allows users to generate new objects by adjusting the new parameters of the trained autoencoder neural network, which outputs procedural model parameters. The output procedural model parameters may be provided to the procedural model to generate the new objects.

Abstract: In one embodiment, payoffs to an actor during the strategic game according to a solution concept is simulated a predetermined number of times. A machine learning tool is used to predict a payoff based on the predetermined number of payoffs according to the solution concept, wherein the machine learning tool is trained with plurality of pairs, each pair comprising a set of simulated payoffs and an actual payoff, wherein the simulated payoffs are predetermined number of payoffs determined in accordance with the solution concept. In another embodiment, a process tree representative of a strategic game involving actors is obtained. The process tree may comprise computational nodes, and leaf nodes which define payoff based on computations of the computational nodes. The process tree is reduced to a game tree representing the strategic game by simulating routes in the process tree.

Abstract: A method for generating hypotheses in a corpus of data comprises selecting a form of ontology; coding the corpus of data based on the form of the ontology; generating ontology space based on coding results and the ontology; transforming the ontology space into a hypothesis space by grouping hypotheses; weighing hypotheses included in the hypothesis space; and applying a science-based sorting algorithm configured to model a science-based treatment of the weighted hypotheses.

Abstract: Data about vehicle movement at a stoplight are collected. A stoplight cycle time is predicted with a probability model. The data are compared to the predicted stoplight cycle time. A noise function is applied to the data to generate noise-applied data. The probability model for the predicted stoplight cycle time is updated by scaling the probability model with the noise-applied data to generate a new probability model. A recommended vehicle operation is provided via a network to at least one vehicle computer based on the predicted stoplight cycle time determined by the new probability model.

Abstract: A computer-implemented method is provided including receiving sensor data from a mobile device corresponding to a first user. A user state of the first user is predicted based on the sensor data. A request is transmitted to the first user to confirm the predicted user state, and a notification is transmitted regarding the predicted user state to a second user responsive to the first user's confirmation of the predicted user state or the first user's failure to respond to the request. A computing system for monitoring and reporting activity of a mobile device is also provided.