Title:
敵対的事例検知システム、方法およびプログラム
Document Type and Number:
Japanese Patent JP7306460
Kind Code:
B2
Abstract:
An adversarial example detection system capable of detecting adversarial examples at a low computational cost is provided. The preparation unit 100 calculates an inverse matrix of a Gram matrix that is used in a process of approximating a deep learner to a Gaussian process. The output distribution calculation unit 222 calculates mean and variance of output values that are numerical values used for class determination for each class by using the inverse matrix of the Gram matrix, for each input observation data. The probabilistic margin calculation unit 223 calculates a probabilistic margin that is an index of variability of the output values based on the mean and variance of the output values, for each input observation data. The adversarial example detection unit 224 detects the adversarial example from the input observation data based on the probabilistic margin calculated for each input observation data.
More Like This:
Inventors:
Kosuke Yoshida
Application Number:
JP2021538522A
Publication Date:
July 11, 2023
Filing Date:
August 02, 2019
Export Citation:
Assignee:
NEC
International Classes:
G06N3/08
Other References:
LEE,Jaehoon et al.,Deep Neural Networks as Gaussian Processes,[online],v3,arXiv,2018年03月03日,pages:1-17,[令和1年10月21日検索],インターネット
淺原 彰規 他,カーネルガウシアンプロセス回帰による時空間分布データ削減方式,情報処理学会論文誌,日本,情報処理学会,2017年01月15日,第58巻 第1号,第2頁-第12頁
淺原 彰規 他,カーネルガウシアンプロセス回帰による時空間分布データ削減方式,情報処理学会論文誌,日本,情報処理学会,2017年01月15日,第58巻 第1号,第2頁-第12頁
Attorney, Agent or Firm:
Rock wall Fuyuki
Masato Shiokawa
Masato Shiokawa