Abstract: Systems and methods for probabilistic semantic sensing in a sensory network are disclosed. The system receives raw sensor data from a plurality of sensors and generates semantic data including sensed events. The system correlates the semantic data based on classifiers to generate aggregations of semantic data. Further, the system analyzes the aggregations of semantic data with a probabilistic engine to produce a corresponding plurality of derived events each of which includes a derived probability. The system generates a first derived event, including a first derived probability, that is generated based on a plurality of probabilities that respectively represent a confidence of an associated semantic datum to enable at least one application to perform a service based on the plurality of derived events.

Abstract: To select the content to be presented to the user, a first latent vector is determined for a content item based on a first object associated with the content item. A second latent vector is determined for the content item based on a second object associated with the content item. A content item vector is then determined based on the first and second latent vectors. Furthermore, a user vector is determined based on interactions of the user with the first set of content objects and the second set of content objects. A score indicative of the likelihood of the user interacting with the content item is determined based on the content item vector and the user vector.

Abstract: One embodiment of an accelerator includes a computing unit; a first memory bank for storing input activations and a second memory bank for storing parameters used in performing computations, the second memory bank configured to store a sufficient amount of the neural network parameters on the computing unit to allow for latency below a specified level with throughput above a specified level. The computing unit includes at least one cell comprising at least one multiply accumulate (“MAC”) operator that receives parameters from the second memory bank and performs computations. The computing unit further includes a first traversal unit that provides a control signal to the first memory bank to cause an input activation to be provided to a data bus accessible by the MAC operator. The computing unit performs computations associated with at least one element of a data array, the one or more computations performed by the MAC operator.

Abstract: Systems and methods are provided for automatically scoring a constructed response. The constructed response is processed to generate a plurality of numerical vectors that is representative of the constructed response. A model is applied to the plurality of numerical vectors. The model includes an input layer configured to receive the plurality of numerical vectors, the input layer being connected to a following layer of the model via a first plurality of connections. Each of the connections has a first weight. An intermediate layer of nodes is configured to receive inputs from an immediately-preceding layer of the model via a second plurality of connections, each of the connections having a second weight. An output layer is connected to the intermediate layer via a third plurality of connections, each of the connections having a third weight. The output layer is configured to generate a score for the constructed response.

Abstract: Techniques described herein may be used to solve a stochastic problem by dividing the stochastic problem into multiple fragments. In some cases, each fragment may be related to a random variable that forms a part of the problem, such that each fragment may produce samples from a probability distribution for that variable. Each fragment of the stochastic problem may then be assigned to a configurable circuit to solve the stochastic fragment. Configurable circuits may be implemented using any suitable combination of hardware and/or software, including using stochastic circuitry. In some embodiments, stochastic circuitry may include a stochastic tile and/or a stochastic memory.

Abstract: A numerical control system detects a state amount indicating a state of an injection operation of an injection molding machine, generates a characteristic amount that characterizes the state of the injection operation from the state amount, and infers an evaluation value of the state of the injection operation from the characteristic amount. The numerical control system detects an abnormal state on the basis of the evaluation value, generates or updates a learning model by machine learning that uses the characteristic amount, and stores the learning model in correlation with a combination of conditions of the injection operation.

Abstract: Embodiments of the present invention provide systems, methods, and computer storage media for facilitating generation of prediction models. In some embodiments, a predetermined number of parameter value sets is identified. Each parameter value set includes a plurality of parameter values that represent corresponding parameters within a time series model. The parameter values can be selected in accordance with stratified sampling to increase a likelihood of prediction accuracy. The parameter value sets are input into a time series model to generate a prediction value in accordance with observed time series data, and the parameter value set resulting in a least amount of prediction error can be selected and used to generate a time series prediction model (ARIMA, AR, MA, ARMA) with corresponding model parameters, such as p, q, and/or k, subsequently used to predict values.

Abstract: A machine learning device which learns a correlation between shipment inspection information obtained by inspecting an object in shipment thereof and operation alarm information issued during operation of the object, includes a state observation unit which observes the shipment inspection information and the operation alarm information; and a learning unit which generates a learning model based on the shipment inspection information and the operation alarm information observed by the state observation unit.

Abstract: In accordance to embodiments, an encoder neural network is configured to receive a one-hot representation of a real text and output a latent representation of the real text generated from the one-hot representation of the real text. A decoder neural network is configured to receive the latent representation of the real text, and output a reconstructed softmax representation of the real text from the latent representation of the real text, the reconstructed softmax representation of the real text is a soft-text. A generator neural network is configured to generate artificial text based on random noise data. A discriminator neural network is configured to receive the soft-text and receive a softmax representation of the artificial text, and output a probability indicating whether the softmax representation of the artificial text received by the discriminator neural network is not from the generator neural network.

Abstract: Systems and methods for understanding (imputing meaning to) multi model data streams may be used in intelligent surveillance and allow a) real-time integration of streaming data from video, audio, infrared and other sensors; b) processing of the results of such integration to obtain understanding of the situation as it unfolds; c) assessing the level of threat inherent in the situation; and d) generating of warning advisories delivered to appropriate recipients as necessary for mitigating the threat. The system generates understanding of the system by creating and manipulating models of the situation as it unfolds. The creation and manipulation involve “neuronal packets” formed in mutually constraining associative networks of four basic types. The process is thermodynamically driven, striving to produce a minimal number of maximally stable models. Obtaining such models is experienced as grasping, or understanding the input steam (objects, their relations and the flow of changes).

Abstract: Embodiments of the present invention provide a computer-implemented method for adaptive residual gradient compression for training of a deep learning neural network (DNN). The method includes obtaining, by a first learner, a current gradient vector for a neural network layer of the DNN, in which the current gradient vector includes gradient weights of parameters of the neural network layer that are calculated from a mini-batch of training data. A current residue vector is generated that includes residual gradient weights for the mini-batch. A compressed current residue vector is generated based on dividing the residual gradient weights of the current residue vector into a plurality of bins of a uniform size and quantizing a subset of the residual gradient weights of one or more bins of the plurality of bins. The compressed current residue vector is then transmitted to a second learner of the plurality of learners or to a parameter server.

Abstract: According to one embodiment, a first set of features is received, where each of the features in the first set being associated with a predetermined category. A bloom filter is applied to the first set of features to generate a second set of features. A neural network model is trained by applying the second set of features to a first layer of nodes of the neural network model to generate an output, the neural network model including a plurality of layers of nodes coupled to each other via a connection. The output of the neural network model is compared with a target value associated with the predetermined category to determine whether the neural network model satisfies a predetermined condition.

Abstract: Subject matter disclosed herein may relate to storage of signals and/or states representative of parameters in a computing device, and may relate more particularly to storage of signals and/or states representative of neural network parameters in a computing device.

Abstract: A unified model for a neural network can be used to predict a particular value, such as a customer value. In various instances, customer value may have particular sub-components. Taking advantage of this fact, a specific learning architecture can be used to predict not just customer value (e.g. a final objective) but also the sub-components of customer value. This allows improved accuracy and reduced error in various embodiments.

Abstract: Disclosed herein are methods, systems, and apparatus, including computer programs encoded on computer storage media, for generating relationship recommendations. One of the methods includes: determining, by a computing device, parameter entities that comprise event entities associated with node entities; constructing, by the computing device, a knowledge graph based on relationships between the parameter entities and predetermined operator entities; extracting, by the computing device, a newly added relationship from the knowledge graph based on an inference rule; and providing, by the computing device, a relationship recommendation corresponding to the newly added relationship.

Abstract: Mechanisms for evaluating an evidential statement in a corpus of evidence are provided. An evidential statement is received for determining a level of confidence in a hypothetical ontological link of an ontology. A source of the evidential statement is identified and a grading of the source of the evidential statement is determined based on a source grading measurement value indicative of a degree of reliability and credibility of the source. An indication of trustworthiness of the evidential statement is generated based on the source grading measurement value. A representation of the indication of trustworthiness of the evidential statement is output in association with the evidential statement.

Abstract: A pseudo-relevance feedback (PRF) system is disclosed that determines an optimized relevance model for a search query by utilizing a posterior relevance model to estimate the likelihood that an initial set of top-K retrieved documents would be retrieved given the posterior relevance model, re-ranking the top-K documents based on their respective estimates of likelihood of retrieval, determining a rank similarity between the initial ranking of the top-K documents and the re-ranking of the top-K documents, updating one or more model parameters of the posterior relevance model based on the rank similarity, and iteratively performing the above process until the rank similarity is maximized, at which point, the optimized relevance model is obtained.

Abstract: A convolution neural network (CNN) model is trained and pruned at a pruning ratio. The model is then trained and pruned one or more times without constraining the model according to any previous pruning step. The pruning ratio may be increased at each iteration until a pruning target is reached. The model may then be trained again with pruned connections masked. The process of pruning, retraining, and adjusting the pruning ratio may also be repeated one or more times with a different pruning target.

Abstract: Embodiments of the invention provide a method comprising maintaining a library of one or more compositional prototypes. Each compositional prototype is associated with a neurosynaptic program. The method further comprises searching the library based on one or more search parameters. At least one compositional prototype satisfying the search parameters is selected. A neurosynaptic network is generated or extended by applying one or more rules associated with the selected compositional prototypes.

Abstract: A processing unit can operate a first recurrent computational model (RCM) to provide first state information and a predicted result value. The processing unit can operating a first network computational model (NCM) to provide respective expectation values of a plurality of actions based at least in part on the first state information. The processing unit can provide an indication of at least one of the plurality of actions, and receive a reference result value, e.g., via a communications interface. The processing unit can train the first RCM based at least in part on the predicted result value and the reference result value to provide a second RCM, and can train the first NCM based at least in part on the first state information and the at least one of the plurality of actions to provide a second NCM.