Abstract: In some aspects, the present disclosure relates to neural language modeling. In one embodiment, a computer-implemented neural network includes a plurality of neural nodes, where each of the neural nodes has a plurality of input weights corresponding to a vector of real numbers. The neural network also includes an input neural node corresponding to a linguistic unit selected from an ordered list of a plurality of linguistic units, and an embedding layer with a plurality of embedding node partitions. Each embedding node partition includes one or more neural nodes. Each of the embedding node partitions corresponds to a position in the ordered list relative to a focus term, is configured to receive an input from an input node, and is configured to generate an output.

Abstract: A method for selecting a reduced number of model neurons in a neural network includes generating a first sparse set of non-zero decoding vectors. Each of the decoding vector is associated with a synapse between a first neuron layer and a second neuron layer. The method further includes implementing the neural network only with selected model neurons in the first neuron layer associated with the non-zero decoding vectors.

Abstract: Recurrent neural networks (RNNs) can be visualized. For example, a processor can receive vectors indicating values of nodes in a gate of a RNN. The values can result from processing data at the gate during a sequence of time steps. The processor can group the nodes into clusters by applying a clustering method to the values of the nodes. The processor can generate a first graphical element visually indicating how the respective values of the nodes in a cluster changed during the sequence of time steps. The processor can also determine a reference value based on multiple values for multiple nodes in the cluster, and generate a second graphical element visually representing how the respective values of the nodes in the cluster each relate to the reference value. The processor can cause a display to output a graphical user interface having the first graphical element and the second graphical element.

Abstract: A system and method can be used to facilitate a strategy for decision making in a fantasy sports league. The method can be used in conjunction with a web or mobile based application. By entering data into the various matrices associated with the application, a customized data set can be created. This data set can then be used with graphical overlays to facilitate future decision making based on created scores attributable to each individual athlete. The end result being a streamlined process that gives one an advantage over others in the fantasy sports league.

Abstract: Systems and methods are provided for selecting training examples to increase the efficiency of supervised active machine learning processes. Training examples for presentation to a user may be selected according to measure of the model's uncertainty in labeling the examples. A number of training examples may be selected to increase efficiency between the user and the processing system by selecting the number of training examples to minimize user downtime in the machine learning process.

Abstract: A target estimator that properly conditions measurement variates in the case of a series of sensor measurements collected against a target, a system model that captures visible and hidden stochastic information including but not limited to target state, target identity, and sensor measurements and that marginalizes measurement failure and a dynamic mixed quadrature expression facilitating real-time implementation of the estimator are presented.

Abstract: Systems, methods, and apparatus, including computer programs encoded on a computer storage medium, for selecting an actions from a set of actions to be performed by an agent interacting with an environment. In one aspect, the system includes a dueling deep neural network. The dueling deep neural network includes a value subnetwork, an advantage subnetwork, and a combining layer. The value subnetwork processes a representation of an observation to generate a value estimate. The advantage subnetwork processes the representation of the observation to generate an advantage estimate for each action in the set of actions. The combining layer combines the value estimate and the respective advantage estimate for each action to generate a respective Q value for the action. The system selects an action to be performed by the agent in response to the observation using the respective Q values for the actions in the set of actions.

Abstract: The present teaching relates to entity linking. In one example, a text string is received. The text string is segmented to obtain a segmentation with a set of one or more segments of the text string. A set of entities are identified, with respect to the one or more segments, from a plurality of entities as linked to the one or more segments. The identifying is in accordance with a probabilistic model based on surface form information associated with the plurality of entities.

Abstract: A computing device automatically classifies an observation vector. (a) A converged classification matrix is computed that defines a label probability for each observation vector. (b) The value of the target variable associated with a maximum label probability value is selected for each observation vector. Each observation vector is assigned to a cluster. A distance value is computed between observation vectors assigned to the same cluster. An average distance value is computed for each observation vector. A predefined number of observation vectors are selected that have minimum values for the average distance value. The supervised data is updated to include the selected observation vectors with the value of the target variable selected in (b). The selected observation vectors are removed from the unlabeled subset. (a) and (b) are repeated. The value of the target variable for each observation vector is output to a labeled dataset.

Abstract: A method described herein includes receiving a digital image, wherein the digital image includes a first element that corresponds to a first domain and a second element that corresponds to a second domain. The method also includes automatically assigning a label to the first element in the digital image based at least in part upon a computed probability that the label corresponds to the first element, wherein the probability is computed through utilization of a first model that is configured to infer labels for elements in the first domain and a second model that is configured to infer labels for elements in the second domain. The first model receives data that identifies learned relationships between elements in the first domain and elements in the second domain, and the probability is computed by the first model based at least in part upon the learned relationships.

Abstract: A system and method for predicting the formulation and processing method and processing parameters for the formation of a biocomposite material is provided. The system and method utilizes the desired properties for the biocomposite material and utilizes these properties m a prediction system to determine the particular formulation, processing method and processing parameters for the formation of a biocomposite material having the desired characteristics. This information is output from the prediction system to a biocomposite material manufacturing system in order to form the biocomposite material and an end product formed therefrom that has the desired characteristics input into the prediction system.

Abstract: A method for performing dataset operations within a cognitive information processing system comprising: receiving data from a plurality of data sources; and, processing the data from the plurality of data sources, the processing the data establishing and maintaining a dynamic data ingestion and enrichment pipeline.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for training and applying a machine learning model. One of the methods includes the actions of obtaining a plurality of data points associated with a specified object; using a machine learning model to generate a prediction from the obtained plurality of data points, the prediction indicating a likelihood that the object will satisfy a particular parameter and a predicted scope for the parameter, wherein the machine learning model is trained using a training set comprising a collection of data points associated with a labeled set of objects, the label indicating the particular parameter and value for each object of the training set; and based on the prediction, classifying the specified object according to a determination of whether the predicted scope satisfies a threshold value.

Abstract: Described is a system for early detection of events via social media mining. The system receives, as input, social media blog posts comprising textual data. The system processes the social media blog posts through a cascade of filters. The cascade of filters comprises an event term detection filter, a location term detection filter following the event term detection filter, and a future date detection filter following the location term detection filter. A plurality of candidate social media blog posts describing an event of interest on a future date is output to a user for further analysis.

Abstract: A circuit element of a multi-dimensional dynamic adaptive neural network array (DANNA) may comprise a neuron/synapse select input functional to select the circuit element to function as one of a neuron and a synapse. In one embodiment of a DANNA array of such circuit elements, (wherein a circuit element may be digital), a destination neuron may be connected to a first neuron by a first synapse in one dimension a second destination neuron may be connected to the first neuron by a second synapse in a second dimension to form linked columns and rows of neuron/synapse circuit elements. In one embodiment, the rows and columns of circuit elements have read registers that are linked together by signal lines and clocked and controlled so as to output columnar data to an output register when a neuron/synapse data value is stored in the read register.

Abstract: A novel method and/or system of feature selection is described.

Abstract: Methods and systems to translate input labels of arcs of a network, corresponding to a sequence of states of the network, to a list of output grammar elements of the arcs, corresponding to a sequence of grammar elements. The network may include a plurality of speech recognition models combined with a weighted finite state machine transducer (WFST). Traversal may include active arc traversal, and may include active arc propagation. Arcs may be processed in parallel, including arcs originating from multiple source states and directed to a common destination state. Self-loops associated with states may be modeled within outgoing arcs of the states, which may reduce synchronization operations. Tasks may be ordered with respect to cache-data locality to associate tasks with processing threads based at least in part on whether another task associated with a corresponding data object was previously assigned to the thread.

Abstract: A system and method includes receiving a tuning work request for tuning an external machine learning model; implementing a plurality of distinct queue worker machines that perform various tuning operations based on the tuning work data of the tuning work request; implementing a plurality of distinct tuning sources that generate values for each of the one or more hyperparameters of the tuning work request; selecting, by one or more queue worker machines of the plurality of distinct queue worker machines, one or more tuning sources of the plurality of distinct tuning sources for tuning the one or more hyperparameters; and using the selected one or more tuning sources to generate one or more suggestions for the one or more hyperparameters, the one or more suggestions comprising values for the one or more hyperparameters of the tuning work request.

Abstract: Automatically detecting and anticipating that an additional machine learning experiment may be needed. A method includes after successfully running a first experiment workflow, automatically prompting a user that an additional experiment workflow may be needed based on specific criteria associated with the first experiment workflow. The method further includes receiving input from the user confirming the additional experiment workflow. As a result of receiving input from the user confirming the additional experiment workflow, the method further includes the system automatically reconfiguring the first experiment workflow, including automatically identifying all necessary modules for the additional experiment workflow and connecting them properly to perform the intended second experiment workflow. The method further includes displaying to the user the first experimental workflow transitioning from the first experiment workflow to the additional experiment workflow.

Abstract: A system and method are disclosed for estimating a number of unique users (e.g., the number of unique users accessing a website, etc.). In one aspect, one or more transactions occurring during a time frame and a plurality of unauthenticated unique identification records associated with the transactions are identified. The time frame is segmented into disjoint time intervals, and a respective bit pattern is determined for each of the unauthenticated unique identification records. A set of churn patterns is determined based on the bit patterns, and a number of expected unauthenticated unique identification records is determined based on the bit patterns and the churn patterns. A number of unique users is estimated based on the size of the set of churn patterns and the number of expected unauthenticated unique identification records.