Abstract: A language proficiency analyzer automatically evaluates a person's language proficiency by analyzing that person's oral communications with another person. The analyzer first enhances the quality of an audio recording of a conversation between the two people using a neural network that automatically detects loss features in the audio and adds those loss features back into the audio. The analyzer then performs a textual and audio analysis on the improved audio. Through textual analysis, the analyzer uses a multi-attention network to determine how focused one person is on the other and how pleased one person is with the other. Through audio analysis, the analyzer uses a neural network to determine how well one person pronounced words during the conversation.

Abstract: An FPGA device receives an input matrix. A first convolutional kernel is determined by performing the exclusive nor operations between the input matrix and a first weight vector. A first binary kernel is determined based on the first convolutional kernel. A first layer feature map is determined by convoluting the input matrix using the first binary kernel. A second convolutional kernel is determined by performing the exclusive nor operations between the first feature map and the second weight vector. A pooled kernel is determined based on the second convolutional kernel. A second binary kernel is determined based on the pooled kernel. A second layer feature map is determined by convoluting the first layer feature map using the second binary kernel. A probability is determined that the input matrix is associated with a predetermined class of images. If the probability is greater than a threshold, classification results are provided.

Abstract: A language proficiency analyzer automatically evaluates a person's language proficiency by analyzing that person's oral communications with another person. The analyzer first enhances the quality of an audio recording of a conversation between the two people using a neural network that automatically detects loss features in the audio and adds those loss features back into the audio. The analyzer then performs a textual and audio analysis on the improved audio. Through textual analysis, the analyzer uses a multi-attention network to determine how focused one person is on the other and how pleased one person is with the other. Through audio analysis, the analyzer uses a neural network to determine how well one person pronounced words during the conversation.

Abstract: Systems, computer program products, and methods are described herein for character recognition in a digital image processing environment. The present invention is configured to electronically retrieve one or more documents from a document repository, wherein the one or more documents are in an image format; initiate one or more image super resolution algorithms on the one or more documents; generate, based on at least the one or more image super resolution algorithms, one or more high-resolution images associated with each of the one or more documents; initiate one or more image bottleneck ensembles (IBE) algorithms on the one or more high-resolution images; extract, using the one or more IBE algorithms, one or more features associated with the one or more high resolution images; and store the one or more features extracted from the one or more high resolution images in a feature repository.

Abstract: Embodiments of the invention are directed to systems, methods, and computer program products for a parameter archival storage system for image processing models. The system is configured for read-optimized compression storage of machine-learning neural-network based image processing models with reduced storage by separately storing weight filter bits. The system is configured to construct weigh parameter objects associated with the plurality of neural network layers of an image processing model, such that the image processing model can be reconstructed from the weigh parameter objects. The system may discard the hierarchical linked architecture of the second image processing model and store the second image processing model at the at least one hosted model versioning system repository by storing only the weigh parameter objects.

Abstract: A device configured to receive an impact report request that comprises a program identifier for a first program. The device is further configured to determine a location within a spatial domain for the first program. The device is further configured to determine a first distance threshold value that corresponds with a first distance away from the location of the first program within the spatial domain. The device is further configured to determine distances between the location of the first program and locations of other programs from the plurality of programs and to identify one or more programs from the plurality of programs that are less than the first distance threshold value. The device is further configured to generate an impact report that identifies the one or more programs from the plurality of programs that are less than the first distance threshold value and to output the generated impact report.

Abstract: A device configured to receive an impact report request that comprises a program identifier for a first program. The device is further configured to determine a location within a spatial domain for the first program. The device is further configured to determine a first distance threshold value that corresponds with a first distance away from the location of the first program within the spatial domain. The device is further configured to determine distances between the location of the first program and locations of other programs from the plurality of programs and to identify one or more programs from the plurality of programs that are less than the first distance threshold value. The device is further configured to generate an impact report that identifies the one or more programs from the plurality of programs that are less than the first distance threshold value and to output the generated impact report.

Abstract: Systems, computer program products, and methods are described herein for generation of data content based on learning reinforcement.

Abstract: Systems, computer program products, and methods are described herein for implementing a cognitive robotic process automation (RPA) architecture. The present invention is configured to electronically receive a video file from a repository, wherein the video file demonstrating one or more actions to be executed in a sequential manner on an application programming interface associated with an application; initiate a neural processing graph generator on the video file; generate, using the neural processing graph generator, a conjugate task graph comprising one or more nodes and one or more edges; initiate a neural task engine on the conjugate task graph; and execute, using the neural task engine, the conjugate task graph.

Abstract: An apparatus is configured to receive an email message that is addressed to a plurality of recipients. The email message comprises a first portion with a first level of encryption and a second portion with a second level of encryption. The apparatus creates a first instance of the message to be sent to a first recipient from among the plurality of recipients. In the first instance, the apparatus masks the portions of the message that are not of the first level of encryption. The apparatus creates a second instance of the message to be sent to a second recipient from among the plurality of recipients. In the second instance, the apparatus masks the portions of the message that are not of the first or second levels of encryption. The apparatus is further configured to transmit the first instance to the first recipient and the second instance to the second recipient.

Abstract: A system includes a production server, a backup server, a telemetry analyzer, a memory, and a hardware processor. The telemetry analyzer takes snapshots of various performance metrics of the production server. The memory stores a log of previous disasters that occurred on the production server. The log includes a snapshot of the production server performance metrics from the time each disaster occurred. The memory also stores recovery scripts for each logged disaster. Each script provides instructions for resolving the linked disaster. The hardware processor uses a machine learning architecture to train an autoencoder. The trained autoencoder receives new snapshots from the telemetry analyzer and generates a reconstruction of the new snapshots. The hardware processor then determines a threshold for distinguishing between server disasters and minor anomalies. This distinction is made by comparing the difference between the reconstruction of the new snapshots and the new snapshots with the threshold.

Abstract: Embodiments of the present invention provide a system for providing a centralized advanced security provisioning platform to create reliable machine learning models and also to enhance the existing machine learning models. The system is configured for executing instructions in the privacy module to monitor and control data privacy and data usage, executing instructions in the security module to preserve the authenticity of data that is used by the machine learning models to predict an outcome, executing instructions in the equality module to detect and prevent biasing of the machine learning models, executing instructions in the transparency module to provide transparency associated with the decision making process employed by the machine learning models, and executing instructions in the accuracy enhancement module to enhance the accuracy of machine learning models.

Abstract: A system for decompiling one or more codes includes a memory operatively coupled to a processor. The system is configured for extracting a binary code and one or more inputs associated with the binary code, disassembling the binary code into a first assembly language code based on the one or more inputs, converting the first assembly language code into first set of fixed dimension vectors, generating a skeleton code from the first set of fixed dimension vectors, optimizing the skeleton code, and generating a high level code.

Abstract: A device configured to obtain project information identifying a set of programs and to generate a program matrix. The device is further configured to identify entries in the program matrix where a row and a column correspond with the same program and to set the identified entries with a value of one. The device is further configured to identify programs from among the set programs that are called by a program and to set the identified entries in the program matrix that correspond with the identified programs with a value of one. The device is further configured to input the program matrix into a first machine learning model to obtain relationship information. The relationship information maps the set of programs to a spatial domain where a distance between a pair of programs in the spatial domain corresponds with a level of impact between the pair of programs.

Abstract: Embodiments of the present invention provide a system for decompiling one or more codes. The system is configured for extracting a binary code and one or more inputs associated with the binary code, disassembling the binary code into a first assembly language code based on the one or more inputs, converting the first assembly language code into first set of fixed dimension vectors, generating a skeleton code from the first set of fixed dimension vectors, optimizing the skeleton code, and generating a high level code.

Abstract: Embodiments of the present invention provide a system for performing automatic code correction for disparate programming languages. The system is configured for identifying defective code lines associated with a code, in response to identifying the defective code lines, extracting the defective code lines, tokenize the defective code lines, passing tokenized defective code lines to an ensemble of neural machine translation models, wherein the ensemble of the neural machine translation models process the tokenized defective code lines, receiving one or more candidates from the ensemble of the neural machine translation models, and generating an output by selecting a candidate from the one or more candidates.

Abstract: A navigation prediction tool tracks and analyzes a user's navigation through a software application to predict the interfaces that are most desired by the user. The tool then dynamically changes the presentation of the software application to present these interfaces to the user. The tool may further conserve resources by predicting how the user will fill certain fields in the interface.

Abstract: Embodiments of the present invention provide a system for generating capsule neural networks for enhancing image processing platforms. The system is configured for generate capsule neural network based on instructions received form at least one user, transfer learning from an existing image processing platform to train the capsule neural network, receive input from one or more devices and provide the input to the existing image processing platform comprising a convolutional neural network, wherein the convolutional neural network processes the input, activate the capsule neural network to validate the processing of the convolutional neural network, and retrain the capsule neural network based on the validations associated with the convolutional neural network.

Abstract: Embodiments of the invention are directed to metamorphic relationship based code testing using mutant generators. The system is configured for identifying and remediating defects in an original program based on constructing at least one mutated program by distorting the original program, and analyzing expression of mutants in test results. In particular, the system receives a request to perform defect analysis of an original program. In response, the system constructs a first mutated program by embedding one or more mutants in the original program code. Moreover, the system typically maps the one or more mutants to one or more metamorphic relationships of process functions of the original program. The system may then implement tests of the original program and the first mutated program by providing one or more predetermined test cases as input to both, to determine whether the original program comprises at least one defect.

Abstract: Embodiments of the invention are directed to systems, methods, and computer program products for an electronic system for management of image processing model database. The system is configured for versioning machine-learning neural-network based image processing models and identifying and tracking mutations in hyper parameters amongst versions of image processing models. The system is configured to determine that a second image processing model is a version of a first image processing model. The system is further configured to map the mutations in hyper parameters between the first plurality of hyper parameters of the first image processing model and the second plurality of hyper parameters associated with the second image processing model.