Naučene indeksne strukture

undergraduate thesis

• Title (english): Task: Learned index structures
• Author: Mate Paulinović
• Mentor: Mile Šikić (mentor)
• Committee member: Mile Šikić (predsjednik povjerenstva)
• Committee member: Krešimir Križanović (član povjerenstva)
• Committee member: Robert Vaser (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2018-09-11, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing

Abstract

Tradicionalne strukture podataka za indeksiranje mogu se poboljšati metodama strojnog učenja. Neuronske mreže za tu primjenu uče distribuciju podataka spremljenih u indeksne strukture. Razvoj grafičkih kartica može povećati efikasnost i smanjiti vremena izvođenja takvih modela.
U ovom radu pokazano je da naučeni model može bolje raspoređivati ključeve po tablici raspršenog adresiranja smanjujuću broj kolizija i memorijsko zauzeće. U radu je predstavljena teorijska podloga tablica raspršenog adresiranja i struktura modela strojnog učenja te implementacijski detalji razvijenog programskog koda.

Abstract (english)

Traditional index data structures can be improved using machine learning.
Neural networks used for this learn the underlying data distributions of the data
which is stored in these data structures. The development of graphics processing
units could improve the efficiency of such models and reduce their execution time.
It is shown in this thesis that a model can be learned which distributes the keys
over the hash table more efficiently, thereby reducing the number of collisions
as well as reducing the memory footprint. This thesis presents the theoretical
background of hash tables and the machine learning structure as well as the
implementation details of the developed source code.

• Keywords: bioinformatika; strojno učenje; ktorka; tablica raspršenog adresiranja; rekurzivni model indeksa; neuronska mreža
• Keywords (english): bioinformatics; machine learning; kmer; hash table; recursive model index; neural network
• Language: croatian
• URN:NBN: urn:nbn:hr:168:253205
• Study programme: Title: Computing; Study programme type: university; Study level: undergraduate; Academic / professional title: sveučilišni/a prvostupnik/ prvostupnica (baccalaureus/baccalaurea) inženjer/inženjerka računarstva (sveučilišni/a prvostupnik/ prvostupnica (baccalaureus/baccalaurea) inženjer/inženjerka računarstva)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Closed access
• Created on: 2019-03-13 23:39:16