Abstract: Machine-learning models (MLM) can be configured more rapidly and accurately according to some examples. For example, a system can receive a first training dataset that includes (i) independent-variable values corresponding to independent variables and (ii) dependent-variable values corresponding to a dependent variable that is influenced by the independent variables. The independent-variable values can include nonlinear-variable values corresponding to at least one nonlinear independent variable. The system can then determine cluster assignments for the nonlinear-variable values, generate a second training dataset based on the cluster assignments, and train a model based on the second training dataset. The trained machine-learning model may then be used in various applications, such as control-system applications.

Abstract: A method of recalibrating a feature data of each channel generated by a convolution layer of a convolution neural network is provided. According to some embodiments, since an affine transformation is applied to the feature data of each channel independently of the feature data of the other channel, the overall number of parameters defining the affine transformation is minimized. As a result, the amount of computations required in performing the feature data recalibration can be reduced.

Abstract: A machine learning device in a component supply device includes: a state observation unit, a determination data acquisition unit, and a learning unit. The state observation unit observes state variables representing a current state of an environment. The state variables include (i) vibration operation parameter data representing an operation parameters for a vibration operation of a tray, (ii) component arrangement data representing an arrangement and a posture of components on the tray, and (iii) component kind data representing a kind of the components. The determination data acquisition unit acquires determination data representing a suitability determination result of the vibration operation, which represents efficiency in supply of the components.

Abstract: Methods, systems, and apparatus for efficiently performing a computation of a convolutional neural network layer. One of the methods includes transforming a X by Y by Z input tensor into a X? by Y? by Z? input tensor; obtaining one or more modified weight matrices, wherein the modified weight matrices operate on the X? by Y? by Z? input tensor to generate a U? by V? by W? output tensor, and the U? by V? by W? output tensor comprises a transformed U by V by W output tensor; and processing the X? by Y? by Z? input tensor using the modified weight matrices to generate the U? by V? by W? output tensor, wherein the U? by V? by W? output tensor comprises the U by V by W output tensor.

Abstract: A method for selecting content to send to labelers for prevalence estimation may include (1) selecting a prevalence estimator, (2) sampling content items from an online system, (3) using, for each of the content items, a model to generate a score for the content item that indicates a likelihood that the content item is of a class of content, (4) generating buckets that each (a) is assigned a range of scores from the model and (b) contains a subset of the content items whose scores fall within the range of scores, (5) determining a sampling rate for each of the buckets that minimizes a variance metric of the estimator, (6) selecting, from each of the buckets, a portion of content items according to the sampling rate of the bucket, and (7) sending the portions to labelers for labeling. Various other methods, systems, and computer-readable media are also disclosed.

Abstract: In some embodiments, a functional object-oriented network (FOON) is provided that includes a plurality of functional units each comprising two or more input object nodes that each identify an object and its state before a manipulation motion is performed, a motion node that identifies a manipulation motion that can be performed using the objects, and one or more output object nodes that each identify an object and its state after the manipulation motion has been performed using the objects. In some embodiment, a robot can used the FOON to determine the discrete actions that are required to perform a given task.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, to present a video. One of the methods includes obtaining one or more unstructured documents. The method includes obtaining, by a computer system, a data model, the data model identifying a type of fact that can be determined from the one or more unstructured documents. The method includes determining, by the computer system, a channel to extract facts from the document based on the type of fact. The method includes distributing, by the computer system, the one or more unstructured documents to the channel. The method includes extracting, by the channel, facts from the one or more unstructured documents. The method also includes storing the facts in a data model.

Abstract: In at least one embodiment, a method and a system for determining a set of plans that best match a set of preferences. The method may include receiving into a goal specification interface at least one goal to be accomplished by the set of plans; receiving into a preference engine a pattern that includes preferences; generating a planning problem by using the preference engine; generating a set of plans by at least one planner; and providing the set of plans for selection of one plan to deploy. In a further embodiment, the preferences may be an occurrence or non-occurrence of at least one component, an occurrence of one component over another component, an ordering between at least two components, an existence or non-existence of at least one tag in a final stream, an existence of one tag over another tag in the final stream.

Abstract: Generally, the present disclosure is directed to systems and methods that perform adaptive optimization with improved convergence properties. The adaptive optimization techniques described herein are useful in various optimization scenarios, including, for example, training a machine-learned model such as, for example, a neural network. In particular, according to one aspect of the present disclosure, a system implementing the adaptive optimization technique can, over a plurality of iterations, employ an adaptive effective learning rate while also ensuring that the effective learning rate is non-increasing.

Abstract: An example method for building a model to predict rare events is disclosed. The example disclosed herein comprises receiving a plurality of historical input logs wherein each log includes at least one key variable and unstructured data. The example further comprises applying text mining techniques to the unstructured data to obtain at least one predictor based on the unstructured data. The example further comprises creating a rare events prediction model based on the at least one key variable and the at least one predictor.

Abstract: A system, method and program product for providing cascade prediction. A system is disclosed having: a computing system for receiving observed cascade data, wherein the observed cascade data includes a set of nodes impacted prior to a preliminary time; a sub-cascade processing engine that determines a sub-cascade size of each node in the set of nodes; survival analysis system that utilizes a networked Weibull regression to determine a survival rate of each node in the set of nodes; and a calculation system that applies the survival rate to the sub-cascade size of each node in the set of nodes to generate a predicted cascade size at a future time.

Abstract: Crowdsourcing systems with machine learning are described. Specifically, item-label inference methods and systems are presented, for example, to provide aggregated answers to a crowdsourced task in a manner achieving good accuracy even where observed data about past behavior of crowd members is sparse. In various examples, an item-label inference system infers variables describing characteristics of both individual crowd workers and communities of the workers. In various examples, an item-label inference system provides aggregated labels while considering the inferred worker characteristics and the inferred characteristics of the worker communities. In examples the item-label inference system provides uncertainty information associated with the inference results for selecting workers and generating future tasks.

Abstract: In some aspects, a quantum computing system includes a multi-dimensional array of qubit devices. Coupler devices reside at intervals between neighboring pairs of the qubit devices in the multi-dimensional array. Each coupler device is configured to produce an electromagnetic interaction between one of the neighboring pairs of qubit devices. In some cases, each qubit device has a respective qubit operating frequency that is independent of an offset electromagnetic field experienced by the qubit device, and the coupling strength of the electromagnetic interaction provided by each coupler device varies with an offset electromagnetic field experienced by the coupler device. In some cases, readout devices are each operably coupled to a single, respective qubit device to produce qubit readout signals that indicate the quantum state of the qubit device.

Abstract: A computer platform implements a precision agriculture system that predicts output conditions, such as diseases, salt damage, soil problems, water leaks and generic anomalies, for orchards under analysis. The computer platform stores site and crop datasets and processed satellite image for the orchards. An orchard data learned model predicts a propensity for existence of output conditions associated with the permanent crops based on the data values for the variables of the site and crop datasets. Also, a satellite model predicts a propensity for existence of the output conditions at the orchard based on processed satellite images. A precision agriculture management model is disclosed that integrates the orchard data learned model with the satellite model to accurately predict the output conditions.

Abstract: The social networking system monitors implicit interactions between a user and objects of the social networking system with which the user has not established a connection. Based on the implicit interactions between the user and an object, the social networking system identifies a soft connection between the user and the object. The social networking system may then identify soft connections to include in a candidate list of soft connections to recommend to the user. The social networking system may also extract signals from the set of candidate list of soft connections, and may use the extracted signals to rank the soft connections in the list of candidate soft connections. The social networking system may then recommend soft connections to the user based on the rank associated with the soft connections in the candidate list of soft connections.

Abstract: A device includes a match element that includes a first data input configured to receive a first result, wherein the first result is of an analysis performed on at least a portion of a data stream by an element of a state machine. The match element also includes a second data input configured to receive a second result, wherein the second result is of an analysis performed on at least a portion of the data stream by another element of the state machine. The match element further includes an output configured to selectively provide the first result or the second result.

Abstract: A system and method for identification of inappropriate multimedia content elements are provided. The method includes receiving a request to identify a multimedia content element from a user device; generating at least one signature respective of the received multimedia content element; matching between the at least one of generated signature respective of the multimedia content element and at least one signature of each concept designated as inappropriate; determining whether a match is identified between the at least one of signature generated respective of the multimedia content element and the at least one signature of an inappropriate concept; and preventing the display on a user device of the multimedia content element, upon identification of a match.

Abstract: Raw data is received from an industrial machine. The industrial machine includes one or more sensors that obtain the data, and the sensors transmit the raw data to a central processing center. The raw data is received at the central processing center and an unsupervised kernel-based algorithm is recursively applied to the raw data. The application of the unsupervised kernel-based algorithm is effective to learn characteristics of the raw data and to determine from the raw data a class of acceptable data. The class of acceptable data is data having a degree of confidence above a predetermined level that the data was obtained during a healthy operation of the machine. The acceptable data is successively determined and refined upon each application of the unsupervised kernel-based algorithm. The unsupervised kernel-based algorithm is executed until a condition is met.

Abstract: Mechanisms for optimizing the determination of supplements for consumers are provided. A longevity inquiry is received from a remote device. A biological extraction is retrieved from a user database and is used with the longevity inquiry to identify a longevity element associated with a user. Further, an ADME model is selected that uses the biological extraction. A machine-learning algorithm is generated using the selected ADME model that uses the longevity element associated with the user as an input and outputs an ADME factor. A tolerant longevity element is identified utilizing the ADME factor.

Abstract: An evaluation platform receives a data set and a description of an outcome, such as predicting results of trends, recognizing patterns, and evaluating options according to specified criteria. The description is evaluated to select candidate evaluators that may be capable of achieving the outcome, and to translate the outcome into a goal for each selected candidate evaluator. The evaluator candidate set is trained using a training data set, and an initial evaluator is selected that exhibits the highest performance to achieve the outcome over the data set. The initial evaluator is applied to achieve the requested outcome over the data set. Optionally, the performance of the initial evaluator may be monitored to detect performance drift. In this event, the evaluator candidate set is reevaluated to identify a substitute evaluator exhibiting higher performance than the initial evaluator, which replaces the initial evaluator in the continued evaluation of the data set.