Efficiency Divide: Comparative Analysis of Human & Neural Network Algorithm Development

Result code in IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F60162694%3AG43__%2F26%3A00563966" target="_blank" >RIV/60162694:G43__/26:00563966 - isvavai.cz</a>

Result on the web

<a href="https://papers.academic-conferences.org/index.php/eccws/article/view/2312" target="_blank" >https://papers.academic-conferences.org/index.php/eccws/article/view/2312</a>

DOI - Digital Object Identifier

<a href="http://dx.doi.org/10.34190/eccws.23.1.2312" target="_blank" >10.34190/eccws.23.1.2312</a>

Result language

angličtina

Original language name

Efficiency Divide: Comparative Analysis of Human & Neural Network Algorithm Development

Original language description

The paper delves into a comparative analysis between human and artificial intelligence (AI) capabilities in algorithm development, with a specific focus on the challenges presented in the "Advent of Code." The research thoroughly investigates the performance of Generative Pre-trained Transformers (GPTs), such as ChatGPT and Bard, in solving intricate algorithmic problems and benchmarks these results against those achieved by human participants. A sizeable portion of the study is dedicated to understanding the nuances of prompt engineering in AI and how it affects the problem-solving process, alongside exploring the choice of programming languages used by both AI and humans. The methodology of the research is extensive, involving the participation of both AI models and human subjects, who vary in their levels of programming expertise. This approach allows for a comprehensive evaluation of the correctness and efficiency of solutions, along with the time taken to resolve the given problems. The results from this study reveal intriguing insights. While AI models like GPTs demonstrate an impressive speed in problem resolution, they often fall short in accuracy when compared to human problem-solvers, particularly in tasks demanding deeper contextual understanding and creative reasoning. Furthermore, the study delves into the impact of time constraints on the effectiveness of problem-solving strategies employed by both AI and humans. It finds that under strict time constraints, AI models can quickly generate solutions, but these solutions may lack the depth and accuracy found in those devised by human participants. This aspect of the research highlights the trade-off between speed and precision in AI-driven problem solving. The research extends its implications beyond mere performance comparison. It suggests the potential for a synergistic approach where the computational efficiency and rapid problem-solving abilities of AI can be effectively combined with the nuanced understanding and creative problem-solving skills inherent in humans. This hybrid approach could redefine the future landscape of programming and algorithm development. The study not only provides a critical analysis of the capabilities of AI in the realm of algorithmic problem-solving but also paves the way for future exploration into the collaborative dynamics of human-AI interaction in programming. It highlights the evolving role of AI in programming and underscores the importance of balancing AI's computational prowess with human creativity and adaptability in solving complex, real-world problems.

Czech name

—

Czech description

—

Type

D - Article in proceedings

CEP classification

—

OECD FORD branch

10201 - Computer sciences, information science, bioinformathics (hardware development to be 2.2, social aspect to be 5.8)

Project

—

Continuities

I - Institucionalni podpora na dlouhodoby koncepcni rozvoj vyzkumne organizace

Publication year

2024

Confidentiality

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Article name in the collection

European Conference on Cyber Warfare and Security

ISBN

978-1-917204-06-4

ISSN

2048-8602

e-ISSN

2048-8610

Number of pages

10

Pages from-to

683-692

Publisher name

ACAD CONFERENCES LTD

Place of publication

—

Event location

Jyvaeskylae, FINLAND

Event date

Jun 27, 2024

Type of event by nationality

WRD - Celosvětová akce

UT code for WoS article

001419654100079