Aspects of the invention include detecting an error alert from a target computer system. In response to detecting the error alert, performance data is then retrieved from the target computer system. A gated recurrent unit (GRU) neural network is used to generate a prediction of a root cause of the error alert based on the performance data. The weights of a reset gate of the GRU neural network are adjusted based on received feedback of the prediction.

Techniques for geometric aging data reduction for machine learning applications are disclosed. In some embodiments, an artificial-intelligence powered system receives a first time-series dataset that tracks at least one metric value over time. The system then generates a second time-series dataset that includes a reduced version of a first portion of the time-series dataset and a non-reduced version of a second portion of the time-series dataset. The second portion of the time-series dataset may include metric values that are more recent than the first portion of the time-series dataset. The system further trains a machine learning model using the second time-series dataset that includes the reduced version of the first portion of the time-series dataset and the non-reduced version of the second portion of the time-series dataset. The trained model may be applied to reduced and/or non-reduced data to detect multivariate anomalies and/or provide other analytic insights.

Techniques for generating imputation-based, uniformly sampled parallel streams of time-series data are disclosed. A system divides into two subsets a dataset made up of multiple data streams. The data streams include interpolated data. The system trains one data correlation model using one subset of the data and applies the trained model to the other subset. The system replaces the interpolated values in the other subset with estimated values generated by the model. The system trains another data correlation model using the revised subset. The system applies the new model to the initial subset to generate estimated values for the initial subset. The system replaces the interpolated values in the initial subset with the estimated values. The system repeats the process of training data correlation models and revising previously-interpolated data points in the subsets of data until a predetermined iteration threshold is met.

A power load data prediction method and device, and a storage medium are disclosed. In an embodiment, the he method comprises: acquiring historical power load data of a one-dimensional time sequence, the historical power load data including values of corresponding time points; mapping the values of corresponding time points to a coordinate system in which a horizontal axis is a set time period, and a vertical axis is time points within the time period, and performing marking at each mapping point by using predetermined pixel values corresponding to the values to obtain a mapping image, wherein different values correspond to different pixel values; and inputting the pixel values of the mapping image to a trained data prediction model, and acquiring a power load data prediction value output by the data prediction model. The method and device and the storage medium can improve the prediction accuracy of the power load data.

A power load data prediction method and device, and a storage medium are disclosed. In an embodiment, the he method comprises: acquiring historical power load data of a one-dimensional time sequence, the historical power load data including values of corresponding time points; mapping the values of corresponding time points to a coordinate system in which a horizontal axis is a set time period, and a vertical axis is time points within the time period, and performing marking at each mapping point by using predetermined pixel values corresponding to the values to obtain a mapping image, wherein different values correspond to different pixel values; and inputting the pixel values of the mapping image to a trained data prediction model, and acquiring a power load data prediction value output by the data prediction model. The method and device and the storage medium can improve the prediction accuracy of the power load data.

Aspects of the invention include determining an event associated with a computing system, the event occurring at a first time, obtaining system data associated with the computing system, determining a system state of the computing system at the first time based on the system data, determining, based on the system state, two or more system data clusters comprising clustered system data associated with the system state of the computing system, determining, via an interpolation algorithm, an interpolated data value for the first time based on the system data, and adjusting the interpolated data value based on a determination that the interpolate data value is outside the two or more system data clusters.

A computer implemented method includes accessing a multivariate time series set of samples collected by multiple biological sensors sensing a first biological function over a first period of time, dividing the data set into windows, calculating statistical dependencies between the samples of the timeseries data collected by each sensor, generating a relationship matrix as a function of the statistical dependencies, and transforming the relationship matrix to generate a first feature vector for each window of time that captures the statistical dependencies amongst the sensors.

Systems, methods, and computer-readable storage media for identifying toxic elements in water, and more specifically to identifying toxins in water based on sensor-detected contaminants and lists of known contaminants. A system can receive water contaminant data from sensors in a predefined geographic area, then normalize that water contaminant data. The system can also receive a list of categorized contaminants and use the list of categorized contaminants and a toxicity of the normalized water contaminants to score water toxicity for that predefined geographic area.

Systems, methods, and computer-readable storage media for identifying toxic elements in water, and more specifically to identifying toxins in water based on sensor-detected contaminants and lists of known contaminants. A system can receive water contaminant data from sensors in a predefined geographic area, then normalize that water contaminant data. The system can also receive a list of categorized contaminants and use the list of categorized contaminants and a toxicity of the normalized water contaminants to score water toxicity for that predefined geographic area.

A system and a method for entity resolution is disclosed. An entity reference parsing subsystem to parse one or more entity references of a corresponding seed set of entity into corresponding one or more personal data properties and property values. A property value standardization subsystem performs one or more standardization operations for standardization of the corresponding one or more property values. A property value anonymization subsystem secures the one or more property values by performing one or more anonymization procedures. A property strength quantification subsystem identifies at least one additional property suspected to belong to the seed set of the entity, assigns a property strength score to the at least one additional property, adds the at least one additional property to the corresponding seed set of entity. A local entity resolution and a global entity resolution subsystem performs a first and a second entity resolution process respectively.