Abstract: Systems, methods, and software for training a machine learning model. The system utilizes training data to train the machine learning model across multiple epochs. The system prepares additional training data by: selecting a set of samples that are unclassified, operating the machine learning model to predict labels that classify the samples, determining an uncertainty of the labels predicted by the machine learning model, calculating a ranking score for each of the samples in the set, selecting a subset of the samples that have more than a threshold ranking score, and submitting the subset to a client for replacement labels. The system receives the replacement labels from the client, and trains the machine learning model, using the subset of the samples as the training data. The labels predicted by the machine learning model for the subset are replaced with corresponding replacement labels from the client.

Abstract: A system and method for iteratively training a neural network are provided. The system and method may include extracting a subset of labeled data points from a pool set of labeled data points; populating, an anchor set of labeled data points with the extracted subset of labeled data points; using the anchor set of labeled data points as labeled inputs to partially train the neural network; selectively swapping, at least some of the labeled data points in the anchor set with at least some of the remaining labeled data points in the pool set; and, retraining the neural network using the anchor set of labeled data points as labeled inputs to the neural network.

Abstract: Method of classifying data may include training, by processing circuitry, a neural network based on labeled inputs of a training data set; identifying, by the processing circuitry, a refinement subset of unlabeled inputs of a pool data set by determining, for each unlabeled input, a first distance of the unlabeled input to the labeled inputs of the training data set and a second distance of the unlabeled input to other unlabeled inputs of the pool data set; submitting, by the processing circuitry, the refinement subset to a labeling process to produce a labeled subset; training, by the processing circuitry, the neural network based on the labeled subset to produce a trained neural network; and classifying, by the processing circuitry, new data using the trained neural network.

Abstract: The present disclosure describes systems and methods for training a neural network. A representative embodiment includes selecting one or more candidate data points to label from a pool set of unlabeled data points where the candidate data point is selected such that the selective addition of the candidate data point and its correct label to a training set of correctly labeled data points used to train the neural network most reduces the label entropy of a test set of unlabeled data points used to test the neural network. The active learning neural network is further (e.g., iteratively) trained using the augmented set of correctly labeled training data points.

Abstract: Systems and methods for regulating data traffic using active learning techniques. In one embodiment, a global monitor communicates with a plurality of sensors over a wide area network. The global monitor builds a global model for a data analytics service that maps elements to values based on data reported by the sensors. The global monitor generates a query for the data from the sensors, selects one or more candidate elements from the elements in the global model, generates a global policy specifying that the data requested from the sensors is limited to the data targeted to the candidate element(s), and sends the query indicating the global policy to the sensors. The global monitor receives the data targeted to the candidate element(s) from the sensors according to the global policy, and adjusts the global model based on the data targeted to the candidate element(s).

Abstract: Systems and methods for regulating data traffic using active learning techniques. In one embodiment, a global monitor communicates with a plurality of sensors over a wide area network. The global monitor builds a global model for a data analytics service that maps elements to values based on data reported by the sensors. The global monitor generates a query for the data from the sensors, selects one or more candidate elements from the elements in the global model, generates a global policy specifying that the data requested from the sensors is limited to the data targeted to the candidate element(s), and sends the query indicating the global policy to the sensors. The global monitor receives the data targeted to the candidate element(s) from the sensors according to the global policy, and adjusts the global model based on the data targeted to the candidate element(s).

Abstract: Various embodiments relate to a network monitor and method thereof for receiving a plurality of alarms during a baseline period from a plurality of network elements, establishing a baseline for at least one of the plurality of the network elements by determining a number of the plurality of alarms during the baseline period that occur during each of a plurality of sub-intervals within the baseline period, calculating a mean arrival rate for each of the plurality of alarms during the baseline period, generating a probability density function for an arrival rate for each of the plurality of alarms during the baseline period, calculating a probabilistic score for each of a plurality of alarms during a test period based on the probability density function for the arrival rate of each of the plurality of alarms during the baseline period, determining if the score for each of the plurality of alarms during the test period is greater than or equal to a probabilistic threshold, summing the number of the scores for ea

Abstract: A capability is provided for adaptive polling of a device based on a set of polling control regions configured to control polling of the device. The set of polling control regions is defined based on at least one of a set of control parameters and non-parametric control information. A transition within the set of polling control regions is determined based on a current polling control region and a target polling control region that is determined based on input information received while in the current polling control region. The input information may include at least one of values of one or more parameters in the set of parameters and non-parametric input information. The transition may include remaining in the current polling control region or transitioning to a new polling control region. The transition may be performed based on a rapid up controlled down (RUCD) transition scheme.

Abstract: Various embodiments relate to a network monitor and method thereof for receiving a plurality of alarms during a baseline period from a plurality of network elements, establishing a baseline for at least one of the plurality of the network elements by determining a number of the plurality of alarms during the baseline period that occur during each of a plurality of sub-intervals within the baseline period, calculating a mean arrival rate for each of the plurality of alarms during the baseline period, generating a probability density function for an arrival rate for each of the plurality of alarms during the baseline period, calculating a probabilistic score for each of a plurality of alarms during a test period based on the probability density function for the arrival rate of each of the plurality of alarms during the baseline period, determining if the score for each of the plurality of alarms during the test period is greater than or equal to a probabilistic threshold, summing the number of the scores for ea

Abstract: A processor accesses a plurality of time series of alarms of a plurality of alarm types that are produced by resources of a network function virtualization (NFV) system. The processor identifies clusters of the plurality of alarm types based on similarities between the plurality of time series and determine causal connections between alarm types in the clusters based on temporal proximity and ordering of the alarm types in the clusters. The processor then stores one or more causality templates representative of the causal connections in a memory.

Abstract: A processor accesses a plurality of time series of alarms of a plurality of alarm types that are produced by resources of a network function virtualization (NFV) system. The processor identifies clusters of the plurality of alarm types based on similarities between the plurality of time series and determine causal connections between alarm types in the clusters based on temporal proximity and ordering of the alarm types in the clusters. The processor then stores one or more causality templates representative of the causal connections in a memory.

Abstract: Embodiments described herein provide improvements to the accuracy of location estimates. In one embodiment, an apparatus stores a first plurality of RF signal strength values that are mapped to known locations, and stores a second plurality of RF signal strength values that are not mapped to known locations. The apparatus estimates locations where the second plurality of RF signal strength values was measured based on the first plurality of RF signal strength values and their corresponding known locations, calculates an uncertainty for the estimated locations, and identifies an estimated location having a highest uncertainty. The apparatus directs a measurement device to measure an RF signal strength value at a known location that corresponds with the estimated location, and adds the RF signal strength value to the first plurality of RF signal strength values to reduce an uncertainty in subsequent estimates of locations within the broadcast area.

Abstract: Various exemplary embodiments relate to a method of determining the root cause of service degradation in a network, the method including determining a window of time; determining one or more abnormal Key Quality Indicators (KQIs) in the window; determining one or more abnormal Key Performance Indicators (KPIs) in the window; calculating a conditional probability that each of one or more KPIs is abnormal when a Key Quality Indicator (KQI) is normal; calculating a conditional probability that the each of one or more KPIs is abnormal when the KQI is abnormal; calculating a score for each KPI based upon a divergence of a Beta distribution for the conditional probability that each of one or more KPIs is abnormal when a KQI is normal and a Beta distribution for the conditional probability that the each of one or more KPIs is abnormal when the KQI is abnormal; and generating a representative root-cause list based upon the score for each KPI.

Abstract: Various exemplary embodiments relate to a method of determining the root cause of service degradation in a network, the method including determining a window of time; determining one or more abnormal Key Quality Indicators (KQIs) in the window; determining one or more abnormal Key Performance Indicators (KPIs) in the window; calculating a conditional probability that each of one or more KPIs is abnormal when a Key Quality Indicator (KQI) is normal; calculating a conditional probability that the each of one or more KPIs is abnormal when the KQI is abnormal; calculating a score for each KPI based upon a divergence of a Beta distribution for the conditional probability that each of one or more KPIs is abnormal when a KQI is normal and a Beta distribution for the conditional probability that the each of one or more KPIs is abnormal when the KQI is abnormal; and generating a representative root-cause list based upon the score for each KPI.

Abstract: An illustrative data classifier device includes data storage and at least one processor configured to operate as a query engine and a passive classifier that is configured to predict classification labels for data. The processor is configured to determine a relationship between the data and training data with associated training classification labels. The processor is also configured to assign a weighted version of at least one of the training classification labels to at least one member of the data based on the determined relationship. An illustrative method of classifying data includes predicting classification labels for data by determining a relationship between the data and training data with associated training classification labels. A weighted version of at least one of the training classification labels is assigned to at least one member of the data based on the determined relationship.

Abstract: A method for deriving a composite tie metric for an edge between nodes of a telecommunication call graph includes receiving descriptive data with original values for descriptive attributes associated with a telecommunication call graph formed by edges between nodes. Each edge relates to two nodes. Each original value relates to an edge and a descriptive attribute forming an edge-attribute pair for the corresponding original value. The descriptive data is stored in a local storage device. Scaling factors for each descriptive attribute are determined taking into account a distribution of the original values for the corresponding descriptive attribute and a common base for the descriptive attributes. Weighting factors are determined for each descriptive attribute. The composite tie metric is computed for an edge based on the original value, scaling factor, and weighting factor for the descriptive attributes.

Abstract: A capability for classifying a device problem of a customer premises device of a customer is presented. The capability for classifying a device problem of a customer premises device of a customer may include receiving customer trouble ticket information of a trouble ticket associated with a customer, receiving customer device measurement information for a customer premises device of the customer, and classifying a device problem of the customer premises device based on text mining of the customer trouble ticket information and based on the customer device measurement information.

Abstract: A capability is provided for adaptive polling of a device based on a set of polling control regions configured to control polling of the device. The set of polling control regions is defined based on at least one of a set of control parameters and non-parametric control information. A transition within the set of polling control regions is determined based on a current polling control region and a target polling control region that is determined based on input information received while in the current polling control region. The input information may include at least one of values of one or more parameters in the set of parameters and non-parametric input information. The transition may include remaining in the current polling control region or transitioning to a new polling control region. The transition may be performed based on a rapid up controlled down (RUCD) transition scheme.

Abstract: A capability is provided for adaptive polling of a device based on a set of polling control regions configured to control polling of the device. The set of polling control regions is defined based on at least one of a set of control parameters and non-parametric control information. A transition within the set of polling control regions is determined based on a current polling control region and a target polling control region that is determined based on input information received while in the current polling control region. The input information may include at least one of values of one or more parameters in the set of parameters and non-parametric input information. The transition may include remaining in the current polling control region or transitioning to a new polling control region. The transition may be performed based on a rapid up controlled down (RUCD) transition scheme.

Abstract: A capability is provided for adaptive polling of a device based on a set of polling control regions configured to control polling of the device. The set of polling control regions is defined based on at least one of a set of control parameters and non-parametric control information. A transition within the set of polling control regions is determined based on a current polling control region and a target polling control region that is determined based on input information received while in the current polling control region. The input information may include at least one of values of one or more parameters in the set of parameters and non-parametric input information. The transition may include remaining in the current polling control region or transitioning to a new polling control region. The transition may be performed based on a rapid up controlled down (RUCD) transition scheme.