Techniques are described for providing explanation information for time series-based predictions made using statistical models, such as linear statistical models, examples of which include various Exponential Smoothing models, Autoregressive Integrated Moving Average (ARIMA) models, and others. For a forecast predicted by a statistical model that has been trained upon and/or fit to a set of historical times series data points, an explanation is generated for the forecast, where the explanation for the forecast includes information indicative of the importance or impact or influence of individual time series data points in the set on the forecast. The explanation for the forecast may be output to a user along with the forecast. This enables the user to have some visibility into why the particular forecast was predicted by the statistical model.

A system employs Graph Prototypical Networks (GPN) for few-shot node classification on attributed networks, and a meta-learning framework trains the system by constructing a pool of semi-supervised node classification tasks to mimic the real test environment. The system is able to perform meta-learning on an attributed network and derive a highly generalizable model for handling the target classification task. The meta-learning framework addresses extraction of meta-knowledge from an attributed network for few-shot node classification, and identification of the informativeness of each labeled instance for building a robust and effective model.

A system for ranking electronic documents based on reference frequency includes a central controller in electronic communication with a document database. The central controller maintains a graphical model of the electronic documents that identifies all references between documents. A weight is automatically calculated and assigned to each reference within the graphical model in order to increase the significance of document references which are based on subject matter relevance and decrease the significance of document references which are based on interpersonal relationships or other meritless factors. Using the weighted graphical model, the central controller is able to automatically identify document clusters having similar subject matter, create a probability matrix for each cluster based on the weighted graphical model, and apply a power iteration to each probability matrix to yield a reference-based ranking of the electronic documents within each cluster.

A method including training predictor machine learning models (MLMs) using a first data set. The trained predictor MLMs are trained to predict classifications of data items in the first data set. The method also includes training confidence MLMs using second classifications, output by the trained predictor MLMs. The method also includes generating an aggregated ranked list of classes based on third classifications output by the trained predictor MLMs and second confidences output by the trained confidence MLMs. The method also includes training an ensemble confidence MLM using the aggregated ranked list of classes to generate a trained ensemble confidence MLM. The trained ensemble confidence MLM is trained to predict a corresponding selected classification for each corresponding data item in a training data set containing second data items similar to the first data items.

A slicer in a material drop ejecting three-dimensional (3D) object printer identifies the positions and local densities for a plurality of infill lines within a perimeter to be formed within a layer of an object to be formed by the printer. The local density of each infill line is filtered and a control law is applied to the filtered local density to identify an error in the local density compared to a target density. This process is performed iteratively until the error is within a predetermined tolerance range about the target local density. The error is used to generate machine ready instructions to operate the 3D object printer to achieve the target density for the infill lines.

Predictive analytics techniques are provided for produce leading indicators of economic activity based on observed pedestrian attributes—e.g., appearance and behavior—and other factors determined from a range of available data sources. A consistent, semantic metadata structure is described as well as a hypothesis generating and testing system capable of generating predictive analytics models in a non-supervised or partially supervised mode.

Computer hardware and/or software that perform the following operations: (i) receiving a data model, the data model including nodes representing types of information and edges representing relationships between the types of information; (ii) generating a set of digital twin replicas, where a digital twin replica of the set of digital twin replicas corresponds to a respective node of the data model; (iii) utilizing the set of digital twin replicas to generate simulated data corresponding to the types of information represented by the nodes of the data model; and (iv) combining the simulated data generated by the set of digital twin replicas into a combined set of simulated data based, at least in part, on the edges of the data model.

A system for detecting clouds and cloud shadows is described. In one approach, clouds and cloud shadows within a remote sensing image are detected through a three step process. In the first stage a high-precision low-recall classifier is used to identify cloud seed pixels within the image. In the second stage, a low-precision high-recall classifier is used to identify potential cloud pixels within the image. Additionally, in the second stage, the cloud seed pixels are grown into the potential cloud pixels to identify clusters of pixels which have a high likelihood of representing clouds. In the third stage, a geometric technique is used to determine pixels which likely represent shadows cast by the clouds identified in the second stage. The clouds identified in the second stage and the shadows identified in the third stage are then exported as a cloud mask and shadow mask of the remote sensing image.

A system for detecting an anomaly in a video of a factory automation scene is disclosed. The system may accept the video; accept a set of training feature vectors derived from spatio-temporal regions of a training video, where a spatio-temporal region is associated with one or multiple training feature vectors; partition the video into multiple sequences of video volumes; produce a sequence of binary difference images for each of the video volumes; count occurrences of each of predetermined patterns of pixels in each binary difference image for each of the video volumes to produce an input feature vector including an input motion feature vector defining a temporal variation of counts of the predetermined patterns for each of the video volumes; produce a set of distances based on the produced input feature vectors and the set of training feature vectors; and detect the anomaly based on the produced set of distances.

A time series forecasting service system is disclosed. The system identifies a set of cross-validation parameters to be used for cross-validating a model to be used for generating a requested forecast. The requested forecast includes a time series dataset and a forecast horizon identifying a number of time steps for which a forecast is to be made using the time series dataset. The system identifies an objective function to be minimized for determining optimal values for the set of cross-validation parameters and identifies constraints for the cross-validation parameters. The system uses an optimization technique to determine the optimal values for the cross-validation parameters. The optimization technique performs processing that determines the optimal values by minimizing the objective function while satisfying the set of constraints. The system uses the optimal values for the cross-validation parameters to perform cross-validation of the model to be used for making the requested forecast.