IEICE Technical Committee Submission System
Conference Paper's Information
Online Proceedings
[Sign in]
Tech. Rep. Archives
 Go Top Page Go Previous   [Japanese] / [English] 

Paper Abstract and Keywords
Presentation 2019-07-24 09:30
Generation of Family Resemblance Inference Rules by Boid Annotation and Labeled-LDA -- A Machine Learning Approach to Integrate Inference Attack Analysis and Covert Channel Attack Analysis --
Kosuke Kurebayashi, Tetsuya Morizumi, Hirotsugu Kinoshita (KU) ISEC2019-42 SITE2019-36 BioX2019-34 HWS2019-37 ICSS2019-40 EMM2019-45
Abstract (in Japanese) (See Japanese page) 
(in English) In this paper we propose a method for machine learning similar chains of words (word chains) similar to "rules for inferring attack and covert channel" using artificial intelligence. The two machine learning elements of this method are Boid annotations (generate a word chain similar to a critical proposition involved in an inference attack with word2vec, and make the word chain a family-like group by the Boid algorithm), and Labeled-LDA (topic analysis of the text linked to the learned critical word chain). These elements are a combination of pointwise mutual information evaluation, KL information evaluation, and average mutual information evaluation repeatedly and repeatedly. That is, a scale on which the transition law is established, a scale on which the perfect law is established, and a scale on which the equivalence relation is established, meaning that the human interprets the text by evaluating the text random variables from different perspectives It is based on the assumption that it is possible to approximate context). In addition, the concept that penetrates objects learned by Boid annotation and L-LDA is based on the similarity relation that does not satisfy the equivalence relation of "Family Resemblance" in the text and word chain. Furthermore, Boid annotations and Labeled-LDA's multi-layered combination learning are used for machine learning of the teaching data itself and machine learning of the text to be evaluated. The critical word chain learned by the teacher text that describes the security model is used as the label of L-LDA for evaluation. By this calculation, a text corresponding to a critical word chain related to an inference attack and a covert channel attack is learned from a collection of texts. The learning result is referred to as an inference rule for covert channel and inference channel analysis learning of access control. In this paper, in the system based on this method, the concrete design of Boid annotation, that is, the syntax analysis of SQLite, word and text, and the concrete design of Word2vec and Boid algorithm using them are also mentioned.
Keyword (in Japanese) (See Japanese page) 
(in English) Artificial intelligence / Machine learning / Access control / Probabilistic security model / Bayesian probabilistic model / LDA / /  
Reference Info. IEICE Tech. Rep., vol. 119, no. 141, SITE2019-36, pp. 243-249, July 2019.
Paper # SITE2019-36 
Date of Issue 2019-07-16 (ISEC, SITE, BioX, HWS, ICSS, EMM) 
ISSN Print edition: ISSN 0913-5685    Online edition: ISSN 2432-6380
Copyright
and
reproduction
All rights are reserved and no part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording, or any information storage and retrieval system, without permission in writing from the publisher. Notwithstanding, instructors are permitted to photocopy isolated articles for noncommercial classroom use without fee. (License No.: 10GA0019/12GB0052/13GB0056/17GB0034/18GB0034)
Download PDF ISEC2019-42 SITE2019-36 BioX2019-34 HWS2019-37 ICSS2019-40 EMM2019-45

Conference Information
Committee ISEC SITE ICSS EMM HWS BioX IPSJ-CSEC IPSJ-SPT 
Conference Date 2019-07-23 - 2019-07-24 
Place (in Japanese) (See Japanese page) 
Place (in English) Kochi University of Technology 
Topics (in Japanese) (See Japanese page) 
Topics (in English) Security, etc. 
Paper Information
Registration To SITE 
Conference Code 2019-07-ISEC-SITE-ICSS-EMM-HWS-BioX-CSEC-SPT 
Language Japanese 
Title (in Japanese) (See Japanese page) 
Sub Title (in Japanese) (See Japanese page) 
Title (in English) Generation of Family Resemblance Inference Rules by Boid Annotation and Labeled-LDA 
Sub Title (in English) A Machine Learning Approach to Integrate Inference Attack Analysis and Covert Channel Attack Analysis 
Keyword(1) Artificial intelligence  
Keyword(2) Machine learning  
Keyword(3) Access control  
Keyword(4) Probabilistic security model  
Keyword(5) Bayesian probabilistic model  
Keyword(6) LDA  
Keyword(7)  
Keyword(8)  
1st Author's Name Kosuke Kurebayashi  
1st Author's Affiliation Kanagawa University (KU)
2nd Author's Name Tetsuya Morizumi  
2nd Author's Affiliation Kanagawa University (KU)
3rd Author's Name Hirotsugu Kinoshita  
3rd Author's Affiliation Kanagawa University (KU)
4th Author's Name  
4th Author's Affiliation ()
5th Author's Name  
5th Author's Affiliation ()
6th Author's Name  
6th Author's Affiliation ()
7th Author's Name  
7th Author's Affiliation ()
8th Author's Name  
8th Author's Affiliation ()
9th Author's Name  
9th Author's Affiliation ()
10th Author's Name  
10th Author's Affiliation ()
11th Author's Name  
11th Author's Affiliation ()
12th Author's Name  
12th Author's Affiliation ()
13th Author's Name  
13th Author's Affiliation ()
14th Author's Name  
14th Author's Affiliation ()
15th Author's Name  
15th Author's Affiliation ()
16th Author's Name  
16th Author's Affiliation ()
17th Author's Name  
17th Author's Affiliation ()
18th Author's Name  
18th Author's Affiliation ()
19th Author's Name  
19th Author's Affiliation ()
20th Author's Name  
20th Author's Affiliation ()
Speaker Author-1 
Date Time 2019-07-24 09:30:00 
Presentation Time 25 minutes 
Registration for SITE 
Paper # ISEC2019-42, SITE2019-36, BioX2019-34, HWS2019-37, ICSS2019-40, EMM2019-45 
Volume (vol) vol.119 
Number (no) no.140(ISEC), no.141(SITE), no.142(BioX), no.143(HWS), no.144(ICSS), no.145(EMM) 
Page pp.243-249 
#Pages
Date of Issue 2019-07-16 (ISEC, SITE, BioX, HWS, ICSS, EMM) 


[Return to Top Page]

[Return to IEICE Web Page]


The Institute of Electronics, Information and Communication Engineers (IEICE), Japan