Abstract: A system can receive first input data indicative of first user input identifying a first diagnostic state of a computing system. The system can, based on the first input data, determine a second diagnostic state of the computing system, the second diagnostic state comprising information of the first diagnostic state. The system can present a first indication of the second diagnostic state via a user interface. The system can receive second input data indicative of second user input confirming the second diagnostic state. The system can determine a recommended action based on the second diagnostic state. The system can present a second indication of the recommended action via the user interface.

Abstract: An apparatus comprises a processing device configured to obtain a first image, to detect one or more designated types of objects in the first image, at least a given one of the designated types of objects comprising at least a given portion of a computing device associated with a service request and, responsive to detecting the given one of the designated types of objects, to identify features of the given portion of the computing device in the first image. The processing device is also configured to perform an augmentation of the identified features of the given portion of the computing device to generate a second image, the second image comprising a modified version of the first image that contains the given portion of the computing device and excludes one or more other portions of the first image, and to process the service request utilizing the second image.

Abstract: One example method includes accessing input data elements from logs that identify user problems with computing system components, the data elements each associated with a respective original class label that identifies a class of computing system components to which the data element relates, the respective original class labels forming a group of class labels, and a first of the original class labels is overrepresented in the group, and reducing overrepresentation of the first original class label in the group by creating an arbitrary aggregation of some of the class labels that includes the first original class label. The method includes creating, based on a hierarchical modeling structure, prepared data in which an original class label is replaced by the aggregation. Next a hierarchical model and benchmark model are trained, and each model generates respective predictions for comparison. An inferencing process is performed to determine which predicted label will be used.

Abstract: A system can receive first input data indicative of first user input identifying a first diagnostic state of a computing system. The system can, based on the first input data, determine a second diagnostic state of the computing system, the second diagnostic state comprising information of the first diagnostic state. The system can present a first indication of the second diagnostic state via a user interface. The system can receive second input data indicative of second user input confirming the second diagnostic state. The system can determine a recommended action based on the second diagnostic state. The system can present a second indication of the recommended action via the user interface.

Abstract: An apparatus comprises a processing device configured to obtain a first image, to detect one or more designated types of objects in the first image, at least a given one of the designated types of objects comprising at least a given portion of a computing device associated with a service request and, responsive to detecting the given one of the designated types of objects, to identify features of the given portion of the computing device in the first image. The processing device is also configured to perform an augmentation of the identified features of the given portion of the computing device to generate a second image, the second image comprising a modified version of the first image that contains the given portion of the computing device and excludes one or more other portions of the first image, and to process the service request utilizing the second image.

Abstract: An apparatus comprises a processing device configured to receive a query to extract information from a document, and to perform two or more iterations of utilizing a machine learning-based information extraction model to extract portions of unstructured text data from the document. In each iteration, a portion of the unstructured text data extracted from the document and an associated relevance score are output. In a first iteration, the query and document are input while in subsequent iterations the query and modified versions of the document are input, the modified versions having previously-extracted portions of the unstructured text data removed therefrom. The processing device is also configured to generate a response to the query comprising a subset of the portions of the unstructured text data extracted from the document determined to have associated relevance scores exceeding a threshold relevance score and at least a threshold level of similarity to the query.

Abstract: An apparatus comprises a processing device configured to receive a query to extract information from a document, and to perform two or more iterations of utilizing a machine learning-based information extraction model to extract portions of unstructured text data from the document. In each iteration, a portion of the unstructured text data extracted from the document and an associated relevance score are output. In a first iteration, the query and document are input while in subsequent iterations the query and modified versions of the document are input, the modified versions having previously-extracted portions of the unstructured text data removed therefrom. The processing device is also configured to generate a response to the query comprising a subset of the portions of the unstructured text data extracted from the document determined to have associated relevance scores exceeding a threshold relevance score and at least a threshold level of similarity to the query.

Abstract: A method includes obtaining information regarding a given asset to be repaired, providing the information regarding the given asset to an encoder of a deep learning model, and receiving, from a decoder of the deep learning model, a recommendation for a troubleshooting action to be performed on the given asset. The method also includes performing the recommended troubleshooting action on the given asset, determining whether the recommended troubleshooting action results in a successful repair of the given asset and, responsive to determining that the recommended troubleshooting action does not result in a successful repair of the given asset, augmenting the information regarding the given asset based at least in part on an output vocabulary of the decoder corresponding to the recommended troubleshooting action. The method further includes repeating the providing, receiving, performing and determining steps utilizing the augmented information regarding the given asset.

Abstract: An apparatus includes at least one processing device configured to obtain information regarding a given asset to be repaired, to generate a recommended troubleshooting action to be performed on the given asset, and to provide the recommended troubleshooting action and the obtained information regarding the given asset as input to an encoder of a machine learning model implementing an attention mechanism. The at least one processing device is also configured to receive, from a decoder of the machine learning model, a predicted success of the recommended troubleshooting action. The at least one processing device is further configured to determine whether the predicted success of the recommended troubleshooting action meets designated criteria, to perform the recommended troubleshooting action responsive to the predicted success meeting the designated criteria, and, to modify the recommended troubleshooting action responsive to the predicted success not meeting the designated criteria.

Abstract: An apparatus includes at least one processing device configured to obtain information regarding a given asset to be repaired, to generate a recommended troubleshooting action to be performed on the given asset, and to provide the recommended troubleshooting action and the obtained information regarding the given asset as input to an encoder of a machine learning model implementing an attention mechanism. The at least one processing device is also configured to receive, from a decoder of the machine learning model, a predicted success of the recommended troubleshooting action. The at least one processing device is further configured to determine whether the predicted success of the recommended troubleshooting action meets designated criteria, to perform the recommended troubleshooting action responsive to the predicted success meeting the designated criteria, and, to modify the recommended troubleshooting action responsive to the predicted success not meeting the designated criteria.

Abstract: A method includes obtaining information regarding a given asset to be repaired, providing the information regarding the given asset to an encoder of a deep learning model, and receiving, from a decoder of the deep learning model, a recommendation for a troubleshooting action to be performed on the given asset. The method also includes performing the recommended troubleshooting action on the given asset, determining whether the recommended troubleshooting action results in a successful repair of the given asset and, responsive to determining that the recommended troubleshooting action does not result in a successful repair of the given asset, augmenting the information regarding the given asset based at least in part on an output vocabulary of the decoder corresponding to the recommended troubleshooting action. The method further includes repeating the providing, receiving, performing and determining steps utilizing the augmented information regarding the given asset.