Vpliv na umetni inteligenci temelječih pomočnikov na pisanje izvorne kode

magistrsko delo

Luka Četina (Author), Luka Pavlič (Mentor)

Abstract

Z napredkom umetne inteligence (UI) postajajo pomočniki za dopolnjevanje kode vse bolj napredni in zmogljivi. V sklopu tega dela smo izvedli sistematičen pregled literature na UI temelječih pomočnikov za dopolnjevanje kode, jih opredelili, predstavili njihovo delovanje in trenutne trende, primerjali glavne funkcionalnosti posameznih pomočnikov ter izpostavili izboljšave, ki jih uporaba UI prinaša. Predstavili smo njihov doprinos k času razvoja ter kakovosti kode. Izvedli smo eksperiment za preverjanje uporabnosti konkretnega pomočnika (Tabnine) pri pisanju kode, uporabniško izkušnjo ter ali bi ga udeleženci priporočili tudi ostalim. Udeleženci so pomočnika ocenili kot zgolj zadovoljivo uporabnega, ocena uporabniške izkušnje je bila v povprečju nevtralna. Večina udeležencev bi pomočnika uporabljala tudi v prihodnje, najverjetneje pa ga ne bi posebej priporočili ostalim. Čeprav razlike zaradi majhnega vzorca niso bile signifikantne, so izkušeni v primerjavi z neizkušenimi pri programiranju, uporabi Jave in ogrodja SpringBoot, pomočnika ocenjevali bolj pozitivno, medtem ko so poznavalci pomočnikov le-tega ocenjevali manj pozitivno od nepoznavalcev.

Keywords

umetna inteligenca;izvorna koda;dopolnjevanje kode;magistrske naloge;

Data

Language:	Slovenian
Year of publishing:	2022
Typology:	2.09 - Master's Thesis
Organization:	UM FERI - Faculty of Electrical Engineering and Computer Science
Publisher:	[L. Četina]
UDC:	004.8:004.415.3(043.2)
COBISS:	113498371
Views:	155
Downloads:	41
Average score:	0 (0 votes)
Metadata:

Other data

Secondary language:	English
Secondary title:	The impact of artificial intelligence-assisted development of source code
Secondary abstract:	The to code completion assistants are getting more advanced and useful as a result of a recent progress in the artificial intelligence (AI). During presented research, we conducted a systematic literature review in the domain of AI-based code completion assistants. It helped us to define them, explain how they work and summarize the current state-of-the-art on the topic. In addition, we also compared the main functionalities of leading assistants. We highlighted the improvements, as a result of AI. We explored their influence on shortened development time, code quality, and conducted an experiment to test their usability, user experience and possible chance of referral. Participants rated the Tabnine assistant’s usability as a satisfactory, while the user experience was on average neutral. Most participants will continue to use the assistant in the future. They will, however, most likely not recommend it to others. Although, the differences were not significant due to the small sample, experienced participants (Java and SpringBoot) rated the assistant more positively than inexperienced ones. Those, familiar with code completion assistants rated it less positively than those unfamiliar with them.
Secondary keywords:	artificial intelligence;source code;code completion;
Type (COBISS):	Master's thesis/paper
Thesis comment:	Univ. v Mariboru, Fak. za elektrotehniko, računalništvo in informatiko, Informatika in tehnologije komuniciranja
Pages:	1 spletni vir (1 datoteka PDF (VII, 71 f.))
ID:	15460785