The disclosure herein relates to AI-based methods and systems of using machine-learning to identify deceptive merchants in payment transactions such as recurring payment transactions. For example, the AI-based systems and methods may train and use an aggregate merchant matcher based on entity matching to identify merchant identifiers and/or acquirers that may be used by a merchant, train and use transaction classifiers to classify transactions as deceptive, recognize merchants based on an N-density aware transaction embedding learned from transaction data, and train and use a merchant classifier to classify merchants as deceptive.

A hyperspectral stereoscopic CubeSat with computer vision and artificial intelligence capabilities consists of a device and a data processing methodology. The device comprises a number of VIS-NIR-TIR hyperspectral sensors, a central processor with memory, a supervisor system running independently of the imager system, radios, a solar panel and battery system, and an active attitude control system. The device is launched into low earth orbit to capture, process, and transmit stereoscopic hyperspectral imagery in the visible and infrared portions of the electromagnetic spectrum. The processing methodology therein comprises computer vision and convolutional neural network algorithms to perform spectral feature identification and data transformations.

Methods, apparatuses, devices and systems for genetic testing and model training are provided. A genetic testing method includes: obtaining genetic data to be processed, an average number of genetic segments corresponding to each position in the genetic data to be processed being less than or equal to a preset threshold; inputting the genetic data to be processed into a feature generation network layer for performing a feature extraction operation to obtain genetic features corresponding to the genetic data to be processed and enhanced features corresponding to the genetic features; and inputting the genetic data to be processed and the enhanced features into a genetic identification network layer for performing a genetic testing operation to obtain a testing result. The present disclosure realizes performing feature extraction operations through low-depth genetic data, obtaining genetic features and enhanced features corresponding to the genetic features, and performing testing operations based on the enhanced features.

Learning latent variable models with deep top-down architectures typically requires inferring latent variables for each training example based on posterior distribution of these latent variables. The inference step relies on either time-consuming long-run Markov chain Monte Carlo (MCMC) sampling or a separate inference model for variational learning. Embodiments of a short-run MCMC, such as a short-run Langevin dynamics, are used herein as an approximate flow-based inference engine. Bias existing in the output distribution of non-convergent short-run Langevin dynamics may be corrected by optimal transport (OT), which aims at transforming the biased distribution produced by finite-step MCMC to the prior distribution with a minimum transport cost. Experiment results verify the effectiveness of the OT correction for the short-run MCMC, and demonstrate that latent variable models trained by the disclosed strategy performed better than the variational auto-encoder in terms of image reconstruction, generation and anomaly detection.

A method, device, and computer program product are designed for non-convolutional image processing in microscopy of an input image into an output image using an artificial neural network with at least one contracting path including layers, at least one expanding path including layers, and at least one filter kernel. The method includes determining, in one or multiple artificial neural network layers, a similarity metric between at least one filter kernel and one output of the previous layer. Additionally, in at least one layer of the contracting path, the resolution of the output of the previous layer is reduced, and, in at least one layer of the expanding path, the resolution of the output of the previous layer is increased. The first artificial neural network layer treats the input image as the output of the previous layer, and the output of the last artificial neural network layer is the output image.

A dual variational autoencoder-generative adversarial network (VAE-GAN) is trained to transform a real video sequence and a simulated video sequence by inputting the real video data into a real video decoder and a real video encoder and inputting the simulated video data into a synthetic video encoder and a synthetic video decoder. Real loss functions and simulated loss functions are determined based on output from a real video discriminator and a simulated video discriminator, respectively. The real loss functions are backpropagated through the real video encoder and the real video decoder to train the real video encoder and the real video decoder based on the real loss functions. The synthetic loss functions are backpropagated through the synthetic video encoder and the synthetic video decoder to train the synthetic video encoder and the synthetic video decoder based on the synthetic loss functions. The real video discriminator and the synthetic video discriminator can be trained to determine an authentic video sequence from a fake video sequence using the real loss functions and the synthetic loss functions. The annotated simulated video can be transformed with the synthetic video encoder and the real video decoder of the dual VAE-GAN to generate a reconstructed annotated real video sequence that includes style elements based on the real video sequence. A second neural network is trained using the reconstructed annotated real video sequence to detect and track objects.

The present disclosure relates to systems, methods, and non-transitory computer-readable media that utilize a neural network having a long short-term memory encoder-decoder architecture to progressively modify a digital image in accordance with a natural language request. For example, in one or more embodiments, the disclosed systems utilize a language-to-operation decoding cell of a language-to-operation neural network to sequentially determine one or more image-modification operations to perform to modify a digital image in accordance with a natural language request. In some cases, the decoding cell determines an image-modification operation to perform partly based on the previously used image-modification operations. The disclosed systems further utilize the decoding cell to determine one or more operation parameters for each selected image-modification operation. The disclosed systems utilize the image-modification operation(s) and operation parameter(s) to modify the digital image (e.g., by generating one or more modified digital images) via the decoding cell.

Techniques are provided to validate a digitized audio signal that is generated by a conference participant. Reference speech features of the conference participant are obtained, either via samples provided explicitly by the participant, or collected passively via prior conferences. The speech features include one or more of word choices, filler words, common grammatical errors, idioms, common phrases, pace of speech, or other features. The reference speech features are compared to features observed in the digitized audio signal. If the reference speech features are sufficiently similar to the observed speech features, the digitized audio signal is validated and the conference participant is allowed to remain in the conference. If the validation is not successful, a variety of possible actions are taken, including alerting an administrator and/or terminating the participant's attendance in the conference.

A method, computer system, and computer program product are provided for mitigating network risk. A plurality of risk reports corresponding to a plurality of network devices in a network are processed to determine a multidimensional risk score for the network. The plurality of risk reports are analyzed using a semantic analysis model to identify one or more factors that contribute to the multidimensional risk score. One or more actions are determined using a trained learning model to mitigate one or more dimensions of the multidimensional risk score. The outcomes of applying the one or more actions are presented to a user to indicate an effect of each of the one or more actions on the multidimensional risk score for the network.

A method for generating one or more emoticons for one or more users with respect to one or more fictional characters is provided. The method includes receiving a first image generated by a multiple localized discriminator (MLD) generative adversarial network (GAN) based on a set of features from multiple sets of features associated with the one or more fictional characters, resulting in generation of an output value associated with each of the plurality of discriminators, determining a weight associated with each of the plurality of discriminators based on a distance between each discriminator and the set of features, generating an image info-graph associated with the first image generated by the MLD GAN upon receiving the first image, calculating a relevance associated with each of the plurality of discriminators based on the image info-graph, the set of features and the distance, and generating a plurality of images representing a plurality of emoticons associated with the one or more fictional characters based on each of the multiple sets of features.