Abstract: A method for training an obfuscation network and a surrogate network is provided. The method includes steps of: a 1st learning device (a) inputting original data of a 1st party, corresponding thereto, into the obfuscation network to generate obfuscated data wherein the 1st party owns the original data or is an entity to whom the original data is delegated; (b) transmitting the obfuscated data and the ground truth to a 2nd learning device corresponding to a 2nd party, and instructing the 2nd learning device to (i) input the obfuscated data into the surrogate network to generate characteristic information, (ii) calculate 1st losses using the ground truth and one of the characteristic information and task specific outputs, and (iii) train the surrogate network minimizing the 1st losses, and transmit the 1st losses to the 1st learning device; and (c) training the obfuscation network minimizing the 1st losses and maximizing 2nd losses.

Abstract: A neural network chip and a related product are provided. The neural network chip (103) includes: a memory (102), a data reading/writing circuit, a convolution calculation circuit, wherein the memory is used for storing a feature map; the data reading/writing circuit is used for reading the feature map from the memory and execute an expansion and zero-padding operation on the feature according to configuration information of the feature map, and sending to the convolution calculation circuit (S401); and the convolution calculation circuit is used for performing convolution calculation on the data obtained after the expansion and zero-padding operation to implement a de-convolution operation (S402). The technical solution has advantages of saving memory usage and bandwidth.

Abstract: Systems and methods for content selection with first and second recommendation engines are disclosed herein. The system can include a memory include a content library database and a model database. The system can include a user device having a first network interface and a first I/O subsystem. The system can include one or more servers that can include a packet selection system and a presentation system. These one or more servers can: receive response data from the user device; provide received response data to a first recommendation engine; alert a second recommendation engine when a selected next node is a placeholder node; retrieve at least one statistical model relevant to selection of next node content; and select next node content based on an output of the at least one statistical model.

Abstract: A method, system, and computer program product for learning a recognition model for recognition processing. The method includes preparing one or more examples for learning, each of which includes an input segment, an additional segment adjacent to the input segment and an assigned label. The input segment and the additional segment are extracted from an original training data. A classification model is trained, using the input segment and the additional segment in the examples, to initialize parameters of the classification model so that extended segments including the input segment and the additional segment are reconstructed from the input segment. Then, the classification model is tuned to predict a target label, using the input segment and the assigned label in the examples, based on the initialized parameters. At least a portion of the obtained classification model is included in the recognition model.

Abstract: Methods for text analysis of medical study data to extract predictive data. Natural language processing is performed on a document in a collection of documents to determine whether the document contains medical model data. In response to determining that the document contains medical model data, content relating to the medical model data in the document is annotated. A first medical model is generated based on the annotations for the identified medical model data and a certainty threshold In response to the certainty threshold meeting a user setting, the first medical model is added to a predictive model for determining a risk score, based on the analyzed data.

Abstract: In an example, one or more member profiles and corresponding elapsed times indicating, for each of the one or more member profiles, how long the corresponding member of a social networking service took to respond to a request for confidential data with a submission of confidential data are obtained. Then a first set of one or more features are extracted from the one or more member profiles. The first set of one or more features and corresponding elapsed times are fed into a machine learning algorithm to train a confidential data response time prediction model to output a predicted time to respond to a request for confidential data for a candidate member profile. A second set of one or more features are obtained from a candidate member profile and fed to the confidential data response time prediction model, outputting the predicted time to respond to a request for confidential data.

Abstract: Some embodiments are directed to systems for authoring predictive models. An embodiment includes a computer system implementing a development environment for generating predictive models. The predictive model authoring tool is configured to perform a modeling operation based on one or more user inputs provided to interface controls of the predictive model authoring tool, determine a modeling context for the modeling operation, log the one or more user inputs, generate a predictive model based on one or more model parameters defined during the modeling operation, link the predictive model to an asset, such that one or more sets of data received from the asset are provided to the predictive model during execution of the predictive model, cause the predictive model to be executed such that the predictive model receives data from the asset, and provide the modeling context, the one or more user inputs, and the one or more model parameters.

Abstract: Methods and apparatus, including computer program products, implementing and using techniques for text analysis of medical study data to extract predictive data. Natural language processing is performed on a document in a collection of documents to determine whether the document contains medical model data. In response to determining that the document contains medical model data, content relating to the medical model data in the document is annotated. A first medical model is generated based on the annotations for the identified medical model data and a certainty threshold In response to the certainty threshold meeting a user setting, the first medical model is added to a predictive model for determining a risk score, based on the analyzed data.

Abstract: Embodiments of the invention provide a method and system for machine failure prediction. The method comprises: identifying a plurality of basic memory depth values based on a machine failure history; ascertaining a basic weight range for each of the plurality of basic memory depth values according to a pre-stored table including a plurality of mappings each mapping between a basic memory depth value and a basic weight range, or a predetermined formula for calculating the basic weight range based on the corresponding basic memory depth value; ascertaining a composite initial weight range by calculating an average weight range of the ascertained basic weight range for each identified basic memory depth value; generating initial weights based on the composite initial weight range; and predicting a future failure using a Back Propagation Through Time (BPTT) trained Recurrent Neural Network (RNN) based on the generated initial weights.

Abstract: Training data from multiple types of sensors and captured in previous capture sessions can be fused within a physics-based tracking framework to train motion priors using different deep learning techniques, such as convolutional neural networks (CNN) and Recurrent Temporal Restricted Boltzmann Machines (RTRBMs). In embodiments employing one or more CNNs, two streams of filters can be used. In those embodiments, one stream of the filters can be used to learn the temporal information and the other stream of the filters can be used to learn spatial information. In embodiments employing one or more RTRBMs, all visible nodes of the RTRBMs can be clamped with values obtained from the training data or data synthesized from the training data. In cases where sensor data is unavailable, the input nodes may be unclamped and the one or more RTRBMs can generate the missing sensor data.

Abstract: A method, system and a computer program product are provided for scoring candidate answers for geographic relevance analyzing an input question to identify one or more first geographic foci of the input question based on geographical contextual information associated with the input question, identifying one or more second geographic foci for a candidate answer generated in response to the input question, and then comparing the first and second geographic foci to generate a geographic relevance score for the candidate answer to the input question.

Abstract: A system includes a windowing module that divides time series data for each metric into portions. Each portion corresponds to a respective window of time. A hash module calculates a hash value for each of the portions for each of the metrics. An identification module compares the hash values for each pair of metrics and, for a selected pair of metrics, counts how many windows of time in which the hash values of the selected pair of metrics are equal. A pair is identified as a candidate pair in response to the count exceeding a threshold. A metric graph module creates a first edge in a graph based on the candidate pair of metrics. Each of the metrics is a node in the graph and direct relationships between each pair of the metrics are edges in the graph. An anomaly combination module analyzes an anomaly condition based on the graph.

Abstract: Methods, systems, and apparatus for accessing a N-dimensional tensor are described. In some implementations, a method includes, for each of one or more first iterations of a first nested loop, performing iterations of a second nested loop that is nested within the first nested loop until a first loop bound for the second nested loop is reached. A number of iterations of the second nested loop for the one or more first iterations of the first nested loop is limited by the first loop bound in response to the second nested loop having a total number of iterations that exceeds a value of a hardware property of the computing system. After a penultimate iteration of the first nested loop has completed, one or more iterations of the second nested loop are performed for a final iteration of the first nested loop until an alternative loop bound is reached.

Abstract: A computer-implemented method of analyzing text documents, includes identifying a relationship in a text document associated with an entity, building a predictive model from training data, in response to said identifying a relationship, wherein the predictive model includes a prediction error, and determining whether to store the identified relationship in memory, based on the prediction error.

Abstract: A “Predictive Controller” operates with any type of controller or user input device to predict user inputs or responses to a current state of an application. A predictive model of the current state of the application is applied to prior user inputs to jointly predict a current user-specific psychological state or profile of the user and a predicted next user response or input. The predicted response or input is provided as the user input to the particular application prior to receiving the actual user input, thereby reducing latency of the response of the application to that actual user input. In addition, a tangible feedback corresponding to the predicted next user input is provided. Further, the predictive capabilities of the Predictive Controller can be applied to locally or remotely hosted instances of the application to reduce latencies associated with user inputs received from any type of controller or user input device.

Abstract: The present disclosure is directed to training, and providing recommendations via, a temporal ensemble of neural networks. The neural networks in the temporal ensemble can be trained at different times. For example, a neural network can be periodically trained using current item interaction data, for example once per day using purchase histories of users of an electronic commerce system. The item interaction data can be split into a more recent group and a less recent group, for example the last two weeks of data and the remainder of the last two years of data. The periodic training of neural networks, using updated data and the sliding windows created by the date split, results in a number of different models for predicting item interaction events. Using a collection of these neural networks together as a temporal ensemble can increase recommendation accuracy without requiring additional hardware for training.

Abstract: A transformed data set corresponding to a machine learning classifier's training data set is generated. Each transformed record contains a modified version of a corresponding training record, as well as the prediction made for the training record by the classifier. A set of explanatory rules is minded from the transformed data set, with each rule indicating a relationship between the prediction and one or more features corresponding to the training records. From among the rule set, a particular matching rule is selected to provide an easy-to-understand explanation for a prediction made by the classifier for an observation record which is not part of the training set.

Abstract: Provided are a circuit configuration optimization apparatus and a machine learning device capable of reducing the occurrence frequency of a malfunction based on one of the current position and the current time of a FPGA device. The circuit configuration optimization apparatus includes: a state data acquisition section that acquires at least one of a current position and current time of the FPGA device as state data; and a circuit configuration determination section that determines a circuit configuration of the FPGA device based on the state data acquired by the state data acquisition section, and outputs a command value for reconfiguring the determined circuit configuration on the FPGA device.

Abstract: Systems and methods are provided for a personal entity modeling for computing devices. For example, a computing device comprises at least one processor and memory storing instructions that, when executed by the at least one processor, cause the mobile device to perform operations including identifying a personal entity in content generated for display on the mobile device, generating training examples for the personal entity from the content, and updating an embedding used to model the personal entity using the training examples. The embedding may be used to make predictions regarding the personal entity. For example, the operations may also include predicting an association between a first personal entity displayed on the computing device and a second entity based on the embedding, and providing a recommendation, to be displayed on the computing device, related to the second entity.

Abstract: A computer system, computer program product, and computer-implemented method for communicating electronic messages over a communication network coupled thereto are provided. The computer system comprises a network interface for receiving messages sent over the network and addressed to a user of the computer system; and computer executable electronic message processing software. The software comprises instructions for directing the computer system to receive a message over the network, and to identify whether a sender of the received electronic message is a human or a machine. The identifying includes first and second phases of operation. The first phase includes an offline phase employing information and activities resident on the computer system. The second phase includes an online phase employing resources remotely accessible over the network.