Invention Grant
- Patent Title: Machine learned decision guidance for alerts originating from monitoring systems
-
Application No.: US16656414Application Date: 2019-10-17
-
Publication No.: US11809304B2Publication Date: 2023-11-07
- Inventor: Vidar V. Vikjord , Jan-Ove Karlberg
- Applicant: Microsoft Technology Licensing, LLC
- Applicant Address: US WA Redmond
- Assignee: Microsoft Technology Licensing, LLC
- Current Assignee: Microsoft Technology Licensing, LLC
- Current Assignee Address: US WA Redmond
- Agency: Holzer Patel Drennan
- Main IPC: G06F11/00
- IPC: G06F11/00 ; G06F11/36 ; G06N5/045 ; G06N20/00 ; G06N7/01

Abstract:
Described herein is a system and method for determining whether a detected issue in a computing system is a bug introduced by a developer or an intermittent issue. When an issue is detected, information about the issue is received. A determination is then made as to whether the issue is a new issue or whether it has been previously detected. When it is determined that the issue is a new issue, information about the issue is stored in a storage device. A feature vector is generated for the issue and is analyzed to determine a probability that the issue is a bug. When it is determined that the issue is a bug, the occurrence of the issue is reported to a user of the system that can correct the bug. Once the bug is corrected, the correction is provided back to the system and is used to train the system.
Public/Granted literature
- US11740998B2 Machine learned decision guidance for alerts originating from monitoring systems Public/Granted day:2023-08-29
Information query