A system for rideshare matching using locality sensitive hashing is disclosed, including at least one rider device and at least one driver device in operable connection with a network. A rideshare application is in operable communication with the network and configured for matching a driver to a rider within a match pool via an artificial intelligence engine operating a locality sensitive hashing module.

A computer-implemented method for training a normalizing flow. The normalizing flow is configured to determine a first output signal characterizing a likelihood or a log-likelihood of an input signal. The normalizing flow includes at least one first layer which includes trainable parameters. A layer input to the first layer is based on the input signal and the first output signal is based on a layer output of the first layer. The training includes: determining at least one training input signal; determining a training output signal for each training input signal using the normalizing flow; determining a first loss value which is based on a likelihood or a log-likelihood of the at least one determined training output signal with respect to a predefined probability distribution; determining an approximation of a gradient of the trainable parameters; updating the trainable parameters of the first layer based on the approximation of the gradient.

A method and a system of distributed estimation using q-diffusion least mean squares (qDiff-LMS) to modify adaptive filter weights in a decentralized wireless sensor network of N nodes is described. The method includes receiving, at each node, k, a local estimate of a previous time instance weight, w.sub.k(i−1), of an adaptive filter of each neighboring node, l, where l=1, 2, . . . , M, combining the local estimates of the previous time instance weights to generate a linear combination of global diffused weights, Ø.sub.k(i−1), measuring, for each node k, an output, y.sub.k(i), of the adaptive filter of the node k, calculating, for each node k, a desired response, d.sub.k(i); generating, for each node k, an estimation error, e.sub.k.sup.CTA(i) by subtracting the output, y.sub.k(i) from the desired response, d.sub.k(i), and updating the global diffused weights by adding a portion of the estimated error to the global diffused weights.

Methods, systems, and apparatus, including computer programs encoded on computer storage media, for training a convolutional neural network using a regularization scheme. One of the methods includes repeatedly performing the following operations: obtaining a kernel of a particular convolutional layer; applying a Fourier transform to the kernel; generating a decomposition using singular-value decomposition (SVD); generating a regularized diagonal matrix; generating a recomposition; applying an inverse Fourier transform to the recomposition; and training the convolutional neural network on training inputs.

A light field microscope may record a raw light field video of a sample. The raw video recording may be decomposed into a non-negative low-rank component and a non-negative sparse component. The low-rank component may correspond to a static portion of the sample, and the sparse component may correspond to a dynamically changing portion of the sample. Volume reconstruction may be performed on the sparse component to generate a three-dimensional video of the sample, with improved spatial resolution. In some cases, the decomposition is calculated by an alternating direction method of multipliers algorithm, with the non-negativity of the sparse component and low-rank component enforced after each iteration. In some cases, the volume reconstruction is calculated by Richardson-Lucy iteration with regularization. The sample may be fluorescent. The fluorescence may be indicative of neural activity in the sample.

The invention relates to a method for person re-identification based on deep model with multi-loss fusion training strategy. The method uses a deep learning technology to perform preprocessing operations such as flipping, clipping, random erasing and style transfer, and then feature extraction is performed through a backbone network model; joint training of a network is performed by fusing a plurality of loss functions. Compared with other deep learning-based person re-identification algorithms, the present invention greatly improves the performance of person re-identification by adopting a plurality of preprocessing modes, the fusion of three loss functions and effective training strategy.

A local connectivity feature transform (LCFT) is applied to binary document images containing text characters, to generate transformed document images which are then input into a bi-directional Long Short Term Memory (LSTM) neural network to perform character/word recognition. The LCFT transformed image is a gray scale image where the pixel values encode local pixel connectivity information of corresponding pixels in the original binary image. The transform is one that provides a unique transform score for every possible shape represented as a 33 block. In one example, the transform is computed using a 33 weight matrix that combines bit coding with a zigzag pattern to assign weights to each element of the 33 block, and by summing up the weights for the non-zero elements of the 33 block shape.

An adaptive characteristic spectral line screening method and system based on atomic emission spectrum are provided, the method includes: using a set characteristic screening optimization method to perform a plurality of optimization rounds of characteristic screening, obtaining an initialized spectral dataset of each round of the characteristic screening and initialized characteristic population genes; obtaining an optimal characteristic population gene of each round by a set analysis method, a fitness function, and an iteration of a genetic algorithm; obtaining an optimized characteristic spectral information set when the plurality of optimization rounds reach set optimization rounds; performing combination statistics and discriminant analyses on the optimized characteristic spectral information set to complete an adaptive characteristic spectral line screening. The disclosure can efficiently and automatically screen out the characteristic spectral lines that meet the analysis requirements in the complex atomic emission spectrum, thus ensuring the effectiveness and accuracy of screening the characteristic spectral lines.

Disclosed are an information enhancing method and an information enhancing system. The information enhancing method includes: sampling information to obtain a multi-view dataset labelled with feature and class; creating a fix function to represent quantity of fixes; creating a view sub-classifier to represent quality of fixes; unifying the quantity of fixes and the quality of fixes to create a quantity-quality balance model, and resolving the quantity-quality balance model to obtain a fixed multi-view dataset; computing weight of each view and weight of the feature of the fixed information; computing information entropy of a fixed labeled sample based on the weight of the view and the weight of the feature; and selecting a labeled sample based on the information entropy and the weights according to a selected generation manner to generate an unlabeled sample, thereby augmenting the sampled information and realizing information enhancement. By fixing and augmenting the sampled information, the disclosure effectively enhances the sampled information and improves application system performance, thereby offering a better guide to system design.

Systems and methods for domain adaptation for structured output via disentangled representations are provided. The system receives a ground truth of a source domain. The ground truth is used in a task loss function for a first convolutional neural network that predicts at least one output based on inputs from the source domain and a target domain. The system clusters the ground truth of the source domain into a predetermined number of clusters, and predicts, via a second convolutional neural network, a structure of label patches. The structure includes an assignment of each of the at least one output of the first convolutional neural network to the predetermined number of clusters. A cluster loss is computed for the predicted structure of label patches, and an adversarial loss function is applied to the predicted structure of label patches to align the source domain and the target domain on a structural level.