Tehnike učenja višestrukosti za povećanje učinkovitosti analize koja koristi sporedna svojstva kriptografskih uređaja

master's thesis

• Title (english): Manifold Learning Techniques for Increased Efficiency of Side-Channel Analysis
• Author: Eugen Vušak
• Mentor: Alan Jović (mentor)
• Committee member: Alan Jović (predsjednik povjerenstva)
• Committee member: Vlado Sruk (član povjerenstva)
• Committee member: Stjepan Groš (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2020-07-13, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing

Abstract

Cilj ovog diplomskog rada je razmotriti primjenu postupaka učenja mnogostrukosti kako bi se povećala učinkovitost analize koja koristi sporedna sredstva kriptografskih uređaja (SCA). U sklopu toga detaljno su opisana polja SCA i smanjenja dimenzionalnosti te za njih specifične metode korištene u tom polju. Iako je naglasak u ovom radu na metodama učenja mnogostrukosti, kao jedna od referentnih metoda uzeta je i vrlo popularna metoda za smanjenje dimenzionalnosti, PCA. Dodatno je korištena hibridna metoda smanjenja dimenzionalnosti koja prije metoda učenja mnogostrukosti izvodi metodu rijetkih slučajnih projekcija kako bi se napravila transformacija u međuprostor. Za konačnu klasifikaciju koristi se metoda slučajnih šuma zbog velike otpornosti na smetnje.

Abstract (english)

This master's thesis aims to study the application of manifold learning techniques with the intent of increasing the efficiency of side-channel analysis (SCA). With this in mind, both SCA and dimensionality reduction fields, along with their respected practices, are described. Even though the emphasis in this paper is on manifold learning, as one of the reference methods, the highly popular PCA method for dimensionality reduction was also considered. Additionally, a hybrid dimensionality reduction method was applied, in which, before the manifold learning method, sparse random projections are performed to transform data into interspace. The random forest method is used for the final classification due to its high resistance to noise.

• Keywords: učenje mnogostrukosti; analiza koja koristi sporedna svojstva (SCA); entropija pogađanja (GE); smanjenje dimenzionalnosti; AES
• Keywords (english): manifold learning; side-channel analysis (SCA); guessing entropy (GE); dimensionality reduction; AES
• Language: croatian
• URN:NBN: urn:nbn:hr:168:313559
• Study programme: Title: Computing; Study programme type: university; Study level: graduate; Academic / professional title: magistar/magistra inženjer/inženjerka računarstva (magistar/magistra inženjer/inženjerka računarstva)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Closed access
• Created on: 2023-01-13 14:17:26