Patent · US Active

Using gradients to detect backdoors in neural networks

US11132444B2 · kind B2 · utility

3Cited by

2References

20Claims

0Family size

Assignee

International Business Machines Corporation · US

Inventors

Wilka Carvalho · Los Angeles, US
Bryant Chen · San Jose, US
Benjamin J. Edwards · Palm Harbor, US
Taesung Lee · White Plains, US
Ian M. Molloy · Chappaqua, US
Jialong Zhang · White Plains, US

Key dates

Filing date	Apr 16, 2018
Grant date	Sep 28, 2021
Priority date	—
Expiry date	Jul 9, 2039

Classification

Technology area (CPC G)Physics
CPC primaryG06F2221/033
WIPO fieldComputer technology
WIPO sectorElectrical engineering

Abstract

Mechanisms are provided for evaluating a trained machine learning model to determine whether the machine learning model has a backdoor trigger. The mechanisms process a test dataset to generate output classifications for the test dataset, and generate, for the test dataset, gradient data indicating a degree of change of elements within the test dataset based on the output generated by processing the test dataset. The mechanisms analyze the gradient data to identify a pattern of elements within the test dataset indicative of a backdoor trigger. The mechanisms generate, in response to the analysis identifying the pattern of elements indicative of a backdoor trigger, an output indicating the existence of the backdoor trigger in the trained machine learning model.

Source: USPTO / EPO open patent data. Objective bibliographic and citation counts.