Invention Grant
- Patent Title: Iterative machine learning based techniques for value-based defect analysis in large data sets
-
Application No.: US17002274Application Date: 2020-08-25
-
Publication No.: US11620558B1Publication Date: 2023-04-04
- Inventor: Yifan Xu , Xue Wang , Marc Wichterich
- Applicant: Amazon Technologies, Inc.
- Applicant Address: US WA Seattle
- Assignee: Amazon Technologies, Inc.
- Current Assignee: Amazon Technologies, Inc.
- Current Assignee Address: US WA Seattle
- Agency: Kowert, Hood, Munyon, Rankin & Goetzel, P.C.
- Agent Robert C. Kowert
- Main IPC: G06N7/00
- IPC: G06N7/00 ; G06Q30/0601 ; G06K9/62 ; G06N20/00

Abstract:
One or more defect analysis iterations are performed on a collection of records. In a given iteration, a defect presence probability is obtained using a machine learning model for a record group. An estimated audit benefit is then assigned to the record group based at least partly on a defect remediation importance score of the record group. An indication of the estimated audit benefit of the record group is provided to an auditor. An audited defect status for the record group, generated by the auditor, is used to initiate one or more automated actions.
Information query
IPC分类:
G | 物理 |
G06 | 计算;推算或计数 |
G06N | 基于特定计算模型的计算机系统 |
G06N7/00 | 基于特定数学模式的计算机系统 |