Analiza in primerjava kakovosti rešitev velikih jezikovnih modelov pri reševanju spletnih programskih problemov

diplomsko delo

Andraž Oset (Author), Aleš Holobar (Mentor)

Abstract

V diplomskem delu smo analizirali učinkovitost in kakovost rešitev velikih jezikovnih modelov pri reševanju spletnih programskih problemov. V uvodnem delu smo podrobno predstavili spletne označevalne jezike, programske jezike ter spletne tehnologije, uporabljene v našem delu, in opisali področje umetne inteligence, s poudarkom na velikih jezikovnih modelih. V osrednjem delu smo izvedli temeljito analizo teh modelov pri reševanju spletnih programskih problemov. Pripravili smo prilagojen nabor vprašanj. Analiza temelji na modelih ChatGPT in Blackbox AI, ki smo ju uporabili za oceno uspešnosti in kakovosti. Jezikovni model ChatGPT je bil malo boljši, saj je pokazal nekoliko višjo raven kakovosti pri reševanju zastavljenih nalog.

Keywords

umetna inteligenca;veliki jezikovni modeli;analiza;spletno programiranje;diplomske naloge;

Data

Language:	Slovenian
Year of publishing:	2024
Typology:	2.11 - Undergraduate Thesis
Organization:	UM FERI - Faculty of Electrical Engineering and Computer Science
Publisher:	[A. Oset]
UDC:	004.8:004.774(043.2)
COBISS:	219982595
Views:	0
Downloads:	3
Average score:	0 (0 votes)
Metadata:

Other data

Secondary language:	English
Secondary title:	Analysis and comparison of the quality of solutions of large language models in solving web programming problems
Secondary abstract:	In the diploma thesis, we analysed the efficiency and quality of solutions provided by large language models in solving web programming problems. In the introductory part, we detailed the web markup languages, programming languages, and web technologies used in our work, and described the field of artificial intelligence, with an emphasis on large language models. In the central part, we conducted a thorough analysis of these models in solving web programming problems. We prepared a tailored set of questions for this purpose. The analysis is based on ChatGPT and Blackbox AI models, which we used to assess performance and quality. ChatGPT showed slightly better performance, demonstrating a higher level of quality in solving the tasks posed.
Secondary keywords:	Artificial intelligence;Large language models;Analysis;Web programming;
Type (COBISS):	Bachelor thesis/paper
Thesis comment:	Univ. v Mariboru, Fak. za elektrotehniko, računalništvo in informatiko, Računalništvo in informacijske tehnologije
Pages:	1 spletni vir (1 datoteka PDF (XIII, 88 f.))
ID:	24682390