Abstract: Various embodiments of a system and methods for audio source separation via multi-scale feature learning are disclosed.

Abstract: Tracking-based motion deblurring via coded exposure is provided. Fast object tracking is useful for a variety of applications in surveillance, autonomous vehicles, and remote sensing. In particular, there is a need to have these algorithms embedded on specialized hardware, such as field-programmable gate arrays (FPGAs) or application-specific integrated circuits (ASICs), to ensure energy-efficient operation while saving on latency, bandwidth, and memory access/storage. In an exemplary aspect, an object tracker is used to track motion of one or more objects in a scene captured by an image sensor. The object tracker is coupled with coded exposure of the image sensor, which modulates photodiodes in the image sensor with a known exposure function (e.g., based on the object tracking). This allows for motion blur to be encoded in a characteristic manner in image data captured by the image sensor. Then, in post-processing, deblurring is performed using a computational algorithm.

Abstract: A system for time series analysis using attention models is disclosed. The system may capture dependencies across different variables through input embedding and may map the order of a sample appearance to a randomized lookup table via positional encoding. The system may capture capturing dependencies within a single sequence through a self-attention mechanism and determine a range of dependency to consider for each position being analyzed. The system may obtain an attention weighting to other positions in the sequence through computation of an inner product and utilize the attention weighting to acquire a vector representation for a position and mask the sequence to enable causality. The system may employ a dense interpolation technique for encoding partial temporal ordering to obtain a single vector representation and a linear layer to obtain logits from the single vector representation. The system may use a type dependent final prediction layer.

Abstract: Various embodiments of a system and associated method for audio source separation based on generative priors trained on individual sources. Through the use of projected gradient descent optimization, the present approach simultaneously searches in the source-specific latent spaces to effectively recover the constituent sources. Though the generative priors can be defined in the time domain directly, it was found that using spectral domain loss functions leads to good-quality source estimates.

Abstract: Various embodiments of systems and methods for attention models with random features for multi-layered graph embeddings are disclosed.

Abstract: A system estimates spectral radius and leverages local updates from neighboring nodes in a wireless network to iteratively update state values of each node in the network and estimate a spectral radius of the network with guaranteed convergence. A method associated with the system method is a distributed method that efficiently converges to an invertible function of the spectral radius based only on local communications of the network for digital communication models in the presence and/or absence of packet loss, as opposed to conventional centralized methods.

Abstract: A system reconfigures a photovoltaic array used in solar energy based on observed shading conditions to determine an optimal topology of the photovoltaic array to maximize power output. Specifically, the system is designed to reconfigure a photovoltaic array when the photovoltaic array is partially shaded. The system uses a neural network model to determine a topology that maximizes power output of the photovoltaic array based on irradiance data obtained from the photovoltaic array.

Abstract: A system for time series analysis using attention models is disclosed. The system may capture dependencies across different variables through input embedding and may map the order of a sample appearance to a randomized lookup table via positional encoding. The system may capture capturing dependencies within a single sequence through a self-attention mechanism and determine a range of dependency to consider for each position being analyzed. The system may obtain an attention weighting to other positions in the sequence through computation of an inner product and utilize the attention weighting to acquire a vector representation for a position and mask the sequence to enable causality. The system may employ a dense interpolation technique for encoding partial temporal ordering to obtain a single vector representation and a linear layer to obtain logits from the single vector representation. The system may use a type dependent final prediction layer.

Abstract: Various embodiments of a cyber-physical system for providing cloud prediction for photovoltaic array control are disclosed herein.

Abstract: Solar array fault detection, classification, and localization using deep neural nets is provided. A fault-identifying neural network uses a cyber-physical system (CPS) approach to fault detection in photovoltaic (PV) arrays. Customized neural network algorithms are deployed in feedforward neural networks for fault detection and identification from monitoring devices that sense data and actuate each individual module in a PV array. This approach improves efficiency by detecting and classifying a wide variety of faults and commonly occurring conditions (e.g., eight faults/conditions concurrently) that affect power output in utility scale PV arrays.

Abstract: Various embodiments for a connection topology reconfiguration technique for photovoltaic (PV) arrays to maximize power output under partial shading and fault conditions using neural networks are disclosed herein.

Abstract: Various embodiments of a system and associated method for detection of COVID-19 and other respiratory diseases through classification of audio samples are disclosed herein. The system utilizes features directly extracted from the coughing audio and develops automated diagnostic tools for COVID-19. In particular, the present application discusses a novel modification of a deep neural network architecture by using log-mel spectrograms of the audio excerpts and by optimizing a combination of binary cross-entropy and focal loss parameters. One embodiment of the system achieved an average validation AUC of 82.23% and a test AUC of 78.3% at a sensitivity of 80.49%.

Abstract: A method including receiving an input data set. The input data set can include one of a feature domain set or a kernel matrix. The method also can include constructing dense embeddings using: (i) Nyström approximations on the input data set when the input data set comprises the kernel matrix, and (ii) clustered Nyström approximations on the input data set when the input data set comprises the feature domain set. The method additionally can include performing representation learning on each of the dense embeddings using a multi-layer fully-connected network for each of the dense embeddings to generate latent representations corresponding to each of the dense embeddings. The method further can include applying a fusion layer to the latent representations corresponding to the dense embeddings to generate a combined representation. The method additionally can include performing classification on the combined representation. Other embodiments of related systems and methods are also disclosed.

Abstract: Various embodiments of systems and methods for attention models with random features for multi-layered graph embeddings are disclosed.

Abstract: Various embodiments of systems and methods for robust max consensus for wireless sensor networks in the presence of additive noise by determining and removing a growth rate estimate from state values of each node in a wireless sensor network are disclosed.

Abstract: Various embodiments of systems and methods for adaptive video subsampling for energy-efficient object detection are disclosed herein.

Abstract: Various embodiments of a system and associated method for audio source separation based on generative priors trained on individual sources. Through the use of projected gradient descent optimization, the present approach simultaneously searches in the source-specific latent spaces to effectively recover the constituent sources. Though the generative priors can be defined in the time domain directly, it was found that using spectral domain loss functions leads to good-quality source estimates.

Abstract: Various embodiments of a system and associated method for identifying and classifying faults in a photovoltaic array using relatively little labeled data are described herein. In particular, the system builds on existing PU classification techniques by addition of a feedback loop that enables classification of limited operational data of a photovoltaic array by expanding a plurality of features within the operational data based on a learned importance of each feature.

Abstract: Various embodiments of a cyber-physical system for providing cloud prediction for photovoltaic array control are disclosed herein.

Abstract: Tracking-based motion deblurring via coded exposure is provided. Fast object tracking is useful for a variety of applications in surveillance, autonomous vehicles, and remote sensing. In particular, there is a need to have these algorithms embedded on specialized hardware, such as field-programmable gate arrays (FPGAs) or application-specific integrated circuits (ASICs), to ensure energy-efficient operation while saving on latency, bandwidth, and memory access/storage. In an exemplary aspect, an object tracker is used to track motion of one or more objects in a scene captured by an image sensor. The object tracker is coupled with coded exposure of the image sensor, which modulates photodiodes in the image sensor with a known exposure function (e.g., based on the object tracking). This allows for motion blur to be encoded in a characteristic manner in image data captured by the image sensor. Then, in post-processing, deblurring is performed using a computational algorithm.