In some aspects, systems and methods for efficiently clustering a large-scale dataset for improving the construction and training of machine-learning models, such as neural network models, are provided. A dataset used for training a neural network model configured can be clustered into a first set of clusters and a second set of clusters. The neural network model can be constructed with a number of nodes in a hidden layer that is based on the number of clusters in the first set of clusters. The neural network can be trained based on training samples selected from the second set of clusters. In some aspects, the trained neural network model can be utilized to satisfy risk assessment queries to compute output risk indicators for target entities. The output risk indicator can be used to control access to one or more interactive computing environments by the target entities.

The present disclosure describes systems, non-transitory computer-readable media, and methods for accurately and efficiently removing objects from digital images taken from a camera viewfinder stream. For example, the disclosed systems access digital images from a camera viewfinder stream in connection with an undesired moving object depicted in the digital images. The disclosed systems generate a temporal window of the digital images concatenated with binary masks indicating the undesired moving object in each digital image. The disclosed systems further utilizes a 3D to 2D generator as part of a 3D to 2D generative adversarial neural network in connection with the temporal window to generate a target digital image with the region associated with the undesired moving object in-painted. In at least one embodiment, the disclosed systems provide the target digital image to a camera viewfinder display to show a user how a future digital photograph will look without the undesired moving object.

Systems and methods for detecting, decoupling and quantifying unresolved signals in trace signal data in the presence of noise with no prior knowledge of the signal characteristics (e.g., signal peak location, intensity and width) of the unresolved signals. The systems and methods are useful for analyzing any trace data signals having one or multiple overlapping constituent signals and particularly useful for analyzing data signals which often contain an unknown number of constituent signals with varying signal characteristics, such as peak location, peak intensity and peak width, and varying resolutions. A general signal model function is assumed for each unknown, constituent signal in the trace signal data. In a first phase, the number of constituent signals and signal characteristics are determined automatically in a parallel fashion by executing multiple simultaneous evaluations iteratively starting with an initial set of trial signals. Making simultaneous evaluations and systematically reducing the number of trial signals allows for convergence to an optimal, final set of signals in a very fast and efficient manner.

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.

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.

Techniques for generating magnetic resonance (MR) images of a subject from MR data obtained by a magnetic resonance imaging (MRI) system, the techniques including: obtaining input MR data obtained by imaging the subject using the MRI system; generating a plurality of transformed input MR data instances by applying a respective first plurality of transformations to the input MR data; generating a plurality of MR images from the plurality of transformed input MR data instances and the input MR data using a non-linear MR image reconstruction technique; generating an ensembled MR image from the plurality of MR images at least in part by: applying a second plurality of transformations to the plurality of MR images to obtain a plurality of transformed MR images; and combining the plurality of transformed MR images to obtain the ensembled MR image; and outputting the ensembled MR image.

A method is disclosed for audio object extraction from an audio content which includes identifying a first set of projection spaces including a first subset for a first channel and a second subset for a second channel of the plurality of channels. The method may further include determining a first set of correlations between the first and second channels, each of the first set of correlations corresponding to one of the first subset of projection spaces and one of the second subset of projection spaces. Still further, the method may include extracting an audio object from an audio signal of the first channel at least in part based on a first correlation among the first set of correlations and the projection space from the first subset corresponding to the first correlation, the first correlation being greater than a first predefined threshold. Corresponding system and computer program products are also disclosed.

Described a system for anomaly detection using anomaly cueing. In operation, an input image having two-dimensional (2D) image mixtures of primary components is reformatted into one-dimensional (1D) input signals. Blind source signal separation is used to separate the 1D input signals into separate output primary components, which are 1D output signals. The 1D output signals are reformatted into 2D spatially independent component output images. The system then calculates all possible pair product images of the 2D spatially independent component output images and corresponding signal-to-noise ratios. A pair product image is selected based on the peak signal-to-noise ratio and thresholded to identify anomalies in the pair product image. Several types of devices can then be controlled based on the identified anomalies in the pair product image.

The techniques described herein improve methods to equip a computing device to conduct automatic speech recognition (“ASR”) in talker-independent multi-talker scenarios. In some examples, permutation invariant training of deep learning models can be used for talker-independent multi-talker scenarios. In some examples, the techniques can determine a permutation-considered assignment between a model's estimate of a source signal and the source signal. In some examples, the techniques can include training the model generating the estimate to minimize a deviation of the permutation-considered assignment. These techniques can be implemented into a neural network's structure itself, solving the label permutation problem that prevented making progress on deep learning based techniques for speech separation. The techniques discussed herein can also include source tracing to trace streams originating from a same source through the frames of a mixed signal.

A reconfigurable sensor monitoring system includes software tunable filters, each of which is programmable to condition one type of analog signal. A processor coupled to the software tunable filters receives each type of analog signal so-conditioned.