Title:
ROBUSTNESS EVALUATION DEVICE, ROBUSTNESS EVALUATION METHOD AND RECORDING MEDIUM
Document Type and Number:
WIPO Patent Application WO/2021/038788
Kind Code:
A1
Abstract:
This robustness evaluation device is provided with: a similarity calculation unit which calculates the similarity between a feature amount of an input to an authentication model and a feature amount of a template; a local Lipschitz constant estimation unit which estimates a local Lipschitz constant of a function, which calculates the similarity between the feature amount of the input to the authentication model and the feature amount of the template, in a sphere with the input to the authentication model as the center; and an evaluation value estimation unit which estimates an evaluation value of the robustness of the authentication model on the basis of the similarity, a determination threshold value for the similarity, and the local Lipschitz constant.
Inventors:
KAKIZAKI KAZUYA (JP)
Application Number:
PCT/JP2019/033890
Publication Date:
March 04, 2021
Filing Date:
August 29, 2019
Export Citation:
Assignee:
NEC CORP (JP)
International Classes:
G06F21/57; G06F21/31; G06N99/00
Foreign References:
| JP2018181157A | 2018-11-15 | |||
| JP2015170317A | 2015-09-28 |
Other References:
CARLINI,NICHOLAS ET AL.: "Towards Evaluating the Robustness of Neural Networks", 2017 IEEE SYMPOSIUM ON SECURITY AND PRIVACY (SP), 22 May 2017 (2017-05-22), pages 39 - 57, XP033108086
PAPERNOT, NICOLAS ET AL.: "The Limitations of Deep Learning in Adversarial Settings", 2016 IEEE EUROPEAN SYMPOSIUM ON SECURITY AND PRIVACY (EUROS&P), 21 March 2016 (2016-03-21), pages 372 - 387, XP032899541
PAPERNOT, NICOLAS ET AL.: "The Limitations of Deep Learning in Adversarial Settings", 2016 IEEE EUROPEAN SYMPOSIUM ON SECURITY AND PRIVACY (EUROS&P), 21 March 2016 (2016-03-21), pages 372 - 387, XP032899541
Attorney, Agent or Firm:
TANAI Sumio et al. (JP)
Download PDF: