Login| Sign Up| Help| Contact|

Patent Searching and Data


Title:
ESTIMATION DEVICE
Document Type and Number:
WIPO Patent Application WO/2023/166564
Kind Code:
A1
Abstract:
An estimation device 400 has: an acquisition unit 421 that acquires a plurality of inference results that are respectively inferred as a result of inputting, to a learning model, multiple pieces of candidate data created on the basis of information indicating an unknown attribute candidate; a calculation unit 422 that for each inference result calculates the distance between the inference result acquired by the acquisition unit 421 and a label corresponding to the candidate data; and an estimation unit 423 that estimates an unknown attribute value in accordance with the result calculated by the calculation unit 422.

Inventors:
ENKHTAIVAN BATNYAM (JP)
TSUCHIDA HIKARU (JP)
ITO KUNIHIRO (JP)
TERANISHI ISAMU (JP)
Application Number:
PCT/JP2022/008617
Publication Date:
September 07, 2023
Filing Date:
March 01, 2022
Export Citation:
Click for automatic bibliography generation   Help
Assignee:
NEC CORP (JP)
International Classes:
G06N20/00
Other References:
中村 和晃, 第141回 知っておきたいキーワード Model Inversion Attack, 映像情報メディア学会誌, 01 May 2021, vol. 75, no. 3, pp. 384-386
YUJI HIGUCHI AND OTHERS: "Training data estimation attack using VAE against classification model", PROCEEDINGS OF THE 2020 CRYPTOGRAPHY AND INFORMATION SECURITY SYMPOSIUM (SCIS2020); JANUARY 28-31, 2020, IEICE, JP, 21 January 2020 (2020-01-21) - 31 January 2020 (2020-01-31), JP, pages 1 - 8, XP009549403
UNE MASASHI: "Research trends and issues regarding security of machine learning systems", FINANCIAL RESEARCH, vol. 38, no. 1, 22 January 2019 (2019-01-22), pages 97 - 123, XP093088996
JUN YAJIMA AND OTHERS: "A study on analysis methods for AI security vulnerabilities hidden in machine learning systems", PROCEEDINGS OF THE 2021 CRYPTOGRAPHY AND INFORMATION SECURITY SYMPOSIUM (SCIS 2021); JANUARY 19-22, 2021, IEICE, JP, 19 January 2021 (2021-01-19), JP, pages 1 - 8, XP009549404
Attorney, Agent or Firm:
BABA, Motohiro et al. (JP)
Download PDF: