Brzo prepoznavanje mikroba uporabom dubokog učenja

master's thesis

• Title (english): Rapid Microbe Detection Using Deep Learning
• Author: Sara Bakić
• Mentor: Mile Šikić (mentor)
• Committee member: Mile Šikić (predsjednik povjerenstva)
• Committee member: Tomislav Petković (član povjerenstva)
• Committee member: Krešimir Križanović (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2021-07-01, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing

Abstract

Mikrobi su mikroskopski organizmi nevidljivi golim okom, ali s vrlo važnom ulogom u svakodnevnom životu. Mogućnost njihovog prepoznavanja i točne klasifikacije osnova je u otkrivanju bolesti, prepisivanju lijekova i održavanju zdravog načina života generalno. Glavni cilj ovoga rada bio je razviti metodu baziranu na dubokom učenju koja na brz i točan način otkriva mikrob na temelju kratkih fragmenata te vrste. Arhitektura mreže bazirana je na novoj arhitekturi prezentiranoj u sklopu obrade prirodnog jezika, pod nazivom transformator, koja je za potrebe detekcije mikroba prilagođena i korištena u svrhe dobivanja sažetih reprezentacija sekvencioniranih fragmenata DNA. Kada su sažete reprezentacije pronađene, različite klasifikacijske metode korištene su za finalnu predikciju vrste mikroba.

Abstract (english)

Microbes are microscopic organisms invisible to the naked eye with a significant role in everyday life. The ability to detect and accurately classify them is essential to discover diseases, prescribe medication, keep a healthy lifestyle in general. The main goal of this thesis is develop a deep learning method that can in a fast and accurate way detect a microbial species from short fragments of that species. The architecture was based on a novel NLP architecture called the Transformer, adapted to the microbe detection task and used to obtain compressed representations of the sequenced DNA fragments. Once the compressed representation were found, different classification methods were applied to output predictions of microbial species.

• Keywords: bioinformatika; detekcija mikroba; duboko učenje; pozornost; transformator; trojni gubitak; klasifikacija
• Keywords (english): bioinformatics; microbe detection; deep learning; attention; transformers; triplet loss; classification
• Language: english
• URN:NBN: urn:nbn:hr:168:935178
• 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 11:42:22