Managing data processing system failures using a predictive model as a controller and hidden knowledge from predictive models

A non-transitory machine-readable medium having instructions stored therein, which when executed by a processor of a data processing system manager, cause the processor of the data processing system manager to perform operations for managing failures of data processing systems, the operations comprising:

• obtaining, by a first inference model hosted by the data processing system manager, an input from a user, wherein the input is associated with a failure prediction generated for a data processing system of the data processing systems;
• identifying, by the first inference model, an intention of the user based, at least, on a content of the input;
• parsing, by the first inference model, the content of the input using the intention to obtain one or more data request parameters;
• using, by the first inference model, the one or more data request parameters to obtain a data request for data stored in a structured knowledge repository;
• obtaining, by the first inference model, one or more structured knowledge attributes that are responsive to the data request for the data;
• generating, by the first inference model, a failure prediction response using the one or more structured knowledge attributes, wherein the failure prediction response comprises the failure prediction and a portion of the one or more structured knowledge attributes that provide for interpretability of the failure prediction;
• and
• providing the failure prediction response to the user;
• causing, by the data processing system manager and after the failure prediction is provided to the user, the data processing system to perform one or more remediation actions to obtain an updated one of the data processing system in which actual occurrence of a failure of the data processing system that is indicated in the failure prediction is prevented,
• wherein prior to obtaining the input, the operations further comprise:
• extracting, using one or more techniques comprising explainable artificial intelligence (AI) techniques, the one or more structured knowledge attributes from an architecture of a second inference model different from the first inference model, the second inference model being a trained machine learning model hosted by the data processing system manager that is configured to generate the failure prediction, the architecture being hidden from the user and controls how the second inference model generates inferences, and the failure prediction being one of the inferences generated by the second inference model;
• and
• storing the one or more structured knowledge attributes extracted from the second inference model into the structured knowledge repository.

A data processing system manager, comprising:

• a processor;
• and
• a memory coupled to the processor to store instructions, which when executed by the processor, cause the processor to perform operations for managing failures of data processing systems, the operations comprising:
• obtaining, by a first inference model hosted by the data processing system manager, an input from a user, wherein the input is associated with a failure prediction generated for a data processing system of the data processing systems;
• identifying, by the first inference model, an intention of the user based, at least, on a content of the input;
• parsing, by the first inference model, the content of the input using the intention to obtain one or more data request parameters;
• using, by the first inference model, the one or more data request parameters to obtain a data request for data stored in a structured knowledge repository;
• obtaining, by the first inference model, one or more structured knowledge attributes that are responsive to the data request for the data;
• generating, by the first inference model, a failure prediction response using the one or more structured knowledge attributes, wherein the failure prediction response comprises the failure prediction and a portion of the one or more structured knowledge attributes that provide for interpretability of the failure prediction;
• providing the failure prediction response to the user;
• and
• causing, by the data processing system manager and after the failure prediction is provided to the user, the data processing system to perform one or more remediation actions to obtain an updated one of the data processing system in which actual occurrence of a failure of the data processing system that is indicated in the failure prediction is prevented,
• wherein prior to obtaining the input, the operations further comprise:
• extracting, using one or more techniques comprising explainable artificial intelligence (AI) techniques, the one or more structured knowledge attributes from an architecture of a second inference model different from the first inference model, the second inference model being a trained machine learning model hosted by the data processing system manager that is configured to generate the failure prediction, the architecture being hidden from the user and controls how the second inference model generates inferences, and the failure prediction being one of the inferences generated by the second inference model;
• and
• storing the one or more structured knowledge attributes extracted from the second inference model into the structured knowledge repository.