Abstract: Disclosed are machine learning techniques directed to training a machine learning model for the combined learning of multiple natural language processing (NLP) tasks. The NLP tasks may be named entity recognition (NER), relation extraction (RE), and assertion detection (AD) tasks. The machine learning model may be a multi-layer transformer model. Training the machine learning model may involve first training the NER module on the NER task, and thereafter training the RE module on the RE task while the AD module is simultaneously trained on the AD task. Training the machine learning model may alternatively involve training the NER module on the NER task concurrently with training the RE module on the RE task and training the AD module on the AD task. The trained machine learning model can predict entities and entity types in newly provided text, along with relations between the entities and assertions associated with the entities.

Abstract: Aspects of the disclosure relate to transliteration of machine interpretable languages. The computing platform may receive a query, formatted in a first format for execution on a first database. The computing platform may translate the query to a second format for execution on a second database by: 1) extracting non-essential parameters from the query to create a query key; 2) storing the non-essential parameters; 3) executing a lookup function on a query library to identify a translated query corresponding to the query key; 4) based on identifying that the query library includes portions of the query key rather than the query key, recursively identify the translated query by nesting the portions of the query key; and 5) input the non-essential parameters into the translated query to create an output query. The computing platform may execute the output query on the second database.

Abstract: Machine learning techniques directed to span prediction for textual data are disclosed. As used herein, span prediction is the process of predicting the possible spans of text that can be assigned to a given entity type of a set of predefined entity types. To this end, a machine learning model can be trained to generate values that indicate the predicted probability that a given span of an identified set of spans within text of interest is appropriate for association with a given entity type of the set of predefined entity types. The predicted probability values may be used to determine whether a given span or spans is associated with a given entity type. The predicted spans can also be scored in some examples.

Abstract: Various embodiments provide for summarization of an interaction, conversation, encounter, and/or the like in at least an abstractive manner. In one example embodiment, a method is provided. The method includes generating, using an encoder-decoder machine learning model, a party-agnostic representation data object for each utterance data object. The method further includes generating an attention graph data object to represent semantic and party-wise relationships between a plurality of utterance data objects. The method further includes modifying, using the attention graph data object, the party-agnostic representation data object for each utterance data object to form a party-wise representation data object for each utterance data object. The method further includes selecting a subset of party-wise representation data objects for each of a plurality of parties.

Abstract: Various embodiments of the present invention provide methods, apparatus, systems, computing devices, computing entities, and/or the like for performing natural language processing operations using a hybrid reason code prediction machine learning framework. Certain embodiments of the present invention utilize systems, methods, and computer program products that perform natural language processing using a hybrid reason code prediction machine learning framework that comprises one or more of the following: (i) a hierarchical transformer machine learning model, (ii) an utterance prediction machine learning model, (iii) an attention distribution generation machine learning model, (iv) an utterance-code pair prediction machine learning model, and (v) a hybrid prediction machine learning model.

Abstract: Aspects of the disclosure relate to transliteration of machine interpretable languages. A computing platform may receive a request to perform a data migration from a first database configured in a first format to a second database configured in a second format. The computing platform may receive, from the client application and at an abstraction layer, a query. Based on identifying that the query is formatted for execution at the second database, the computing platform may route the query to the second database for execution. Based on identifying that the query is not formatted for execution at the second database, the computing platform may: 1) translate the query from the first format to the second format by using pre-verified query keys to convert the query from the first format into the second format, and 2) route the translated query to the second database for execution.

Abstract: Aspects of the disclosure relate to transliteration of machine interpretable languages. A computing platform receive a source query formatted in a first format for execution on a source database. The computing platform may execute the source query on the source database to produce a first data result. The computing platform may input the first data result into a reversal logic engine to produce a target query formatted in a second format corresponding to a target database. The computing platform may execute the target query on the target database to produce a second data result. Based on identifying that the second data result matches the first data result, the computing platform may validate the target query. Based on identifying that the second data result does not match the first data result, the computing platform may adjust the reversal logic engine based on the discrepancy.

Abstract: A computing platform may configure a rules-based state machine to predict system failure for a system based on telemetry state images and transitions between the telemetry state images. The computing platform may receive initial telemetry data. The computing platform may generate, based on the initial telemetry data, an initial telemetry state image. The computing platform may receive additional telemetry data, and may generate, based on the additional telemetry data, an additional telemetry state image. The computing platform may compare a pattern, corresponding to the initial telemetry state image, the additional telemetry state image, and a corresponding transition, to historical patterns to identify a match. The computing platform may identify, using the identified matching pattern, a likelihood of failure for the system, and may send, based on the likelihood of failure for the system, preemptive resolution commands causing modification of operations at the system to prevent a predicted failure.

Abstract: A computing platform may train a hybrid deep learning model, including a CNN and RNN, to predict system failure for a system based on telemetry state images and transitions between the telemetry state images. The computing platform may receive initial telemetry data, and may generate an initial telemetry state image. The computing platform may receive additional telemetry data, and may generate an additional telemetry state image. The computing platform may classify, using the CNN and based on historical telemetry state images, the initial telemetry state image and the additional telemetry state image. The computing platform may identify, using the RNN and based on the classified telemetry state images and transitions between the classified telemetry state images, a matching pattern. The computing platform may identify, using the identified matching pattern, a likelihood of failure for the system, and may cause modification of operations at the system to prevent a predicted failure.

Abstract: A computing platform may train, using historical telemetry state images, an image comparison model to identify matches between telemetry state images. The computing platform may generate a plurality of system alerts corresponding to a period of time. The computing platform may access telemetry data corresponding to the period of time. The computing platform may generate, based on the telemetry data and for a time corresponding to each of the plurality of system alerts, a telemetry state image. The computing platform may input, into the image comparison model, the telemetry state images to identify whether or not any of the plurality of telemetry state images match. Based on detecting a match, the computing platform may consolidate system alerts corresponding to the matching telemetry state images, which may produce a single system alert and may send, to a user device, the single system alert.

Abstract: Aspects of the disclosure relate to transliteration of machine interpretable languages. A computing platform may train a machine learning model using source syntax trees and target dialect syntax trees, which may configure the model to output source dialect keys and their corresponding target dialect queries. The computing platform may execute the corresponding target dialect queries to identify whether they are valid. For a valid target dialect query, the computing platform may store the valid target dialect query and first source dialect keys corresponding to the valid target dialect query in a lookup table. For an invalid target dialect query resulting in error, the computing platform may: 1) identify a cause of the error; 2) generate a transliteration rule to correct the error; and 3) store, in the lookup table, the invalid target dialect query, second source dialect keys corresponding to the invalid target dialect query, and the transliteration rule.

Abstract: Various embodiments of the present invention provide methods, apparatus, systems, computing devices, computing entities, and/or the like for performing natural language processing operations using an attention-based text encoder machine learning model that is trained using a multi-task training routine that is associated with two or more training tasks (e.g., a multi-task training routine that is associated with two or more sequential training tasks, a multi-training routine that is associated with two or more concurrent training tasks, and/or the like).

Abstract: There is a need for more effective and efficient text categorization. This need can be addressed by, for example, techniques for semantic text categorization. In one example, a method includes determining an input vector-based representation of an input document; processing the input vector-based representation using a trained supervised machine learning model to generate the categorization based at least in part on the input vector-based representation, wherein: (i) the trained supervised machine learning model has been trained using automatically-generated training data, and (ii) the automatically generated training data is generated by determining an inferred semantic label for each unlabeled training document of one or more unlabeled training documents; and performing one or more categorization-based actions based at least in part on the categorization, and (iii) the labels are described by one or more short documents/short texts.

Abstract: Methods and systems for automatically determining correspondences between communication ports of a networked device and encoders and decoders connected to those communication ports. In some embodiments, the networked device and the encoders and decoders are connected to a video communications network provided by a switch. The networked device can query the video communications network for information related to the encoders and decoders to determine and save the port-to-device correspondences. In some embodiments, the networked device can extract device information from video signals received at its input ports to map the input ports to respectively connected decoders. In similar fashion, the networked device may transmit or embed port-specific information from its output ports to respectively connected encoders. Then, the networked device can query the video communications network for the port-specific information received at the encoders to map the output ports to respectively connected encoders.

Abstract: Aspects of the disclosure relate to transliteration of machine interpretable languages. A computing platform may configure a client application to use a custom driver when communicating with an enterprise database. The computing platform may receive a database query formatted in a first database format corresponding to a first database. The computing platform may translate, using a query translation library, the database query from the first database format into a second database format corresponding to a second database, which may cause the custom driver to execute a transliteration process using pre-verified query keys stored in the query translation library to convert the database query from the first database format into the second database format. The computing platform may execute the translated database query on the second database to obtain a query result, and may send the query result to the client application.

Abstract: Various embodiments of the present disclosure provide methods, apparatus, systems, computing devices, computing entities, and/or the like for generating a representative embeddings for a plurality of temporal sequences by using a graph attention augmented temporal network based on dynamic co-occurrence graphs for preceding temporal sequences and initial embeddings, where the dynamic co-occurrence graphs are projections of a global guidance co-occurrence graph on classification features of the preceding temporal sequences, and the initial embeddings are generated by processing a latent representation of corresponding classification features that is generated by a sequential long short term memory model as well as a classification feature tree using a tree-based long short term memory model.

Abstract: There is a need for more effective and efficient predictive natural language topic detection. This need can be addressed by, for example, solutions for performing sequential topic detection. In one example, a method includes determining a sequential topic distribution data object for the current document sequence, determining a current term-context correlation data object for the current document sequence, determining a current context-topic correlation data object for the current document sequence, determining an updated term-topic correlation data object based at least in part on the current context-topic correlation data object, determining topic modeling predictions based at least in part on the sequential topic distribution data object and the updated term-topic correlation data object, and performing prediction-based actions based at least in part on the topic modeling predictions.

Abstract: Systems, computer program products, and methods are described herein for analyzing system health of individual electronic components using component relational graphs. The method includes receiving a process request. The process request is a request to execute a process. The method also includes determining one or more components of the system used during the process. The method further includes generating a component knowledge graph for the process. The component knowledge graph includes one or more nodes corresponding to each of one or more components used during the process. The method still further includes determining a component health rating for each of the one or more components used in the process.

Abstract: Systems, computer program products, and methods are described herein for analyzing system health of individual electronic components using image mapping. The method includes receiving a component health image for a component based on an execution of a process. The method also includes comparing the component health image based on the execution of the process to previous component health image(s) for the component based on one or more previous executions of the process The method further includes determining a component health image similarity score based on the comparison of the component health image to the one or more previous component health images for the component. The method still further includes determining a component health action based on the component health image similarity score. The component health action includes causing a transmission of an alert in an instance in which the component health image similarity score is outside of a threshold range.

Abstract: Embodiments of the present invention provide a system for analyzing operational parameters of electronic and software components associated with entity applications to detect anomalies. The system is configured for extracting one or more historical images associated with resiliency status of electronic and software components associated with an entity application, analyzing the one or more historical images to generate a pixel wise average of the one or more historical images, generating similarity scores between the one or more historical images, determining a distribution of the similarity scores, receiving a real-time image associated with a current resiliency status of the electronic and software components associated with the entity application, generating a real-time image similarity score for the real-time image, and comparing the real-time image similarity score with the distribution to detect presence of an anomaly.