Context Consistency In AI Dialogue Systems

Evgeny Kren

Bakalářská práce

Abstract:

The purpose of this thesis is to evaluate and compare the abilities of selected AI dialogue systems to maintain context consistency, an essential aspect for effective and engaging human-computer interactions. This thesis will contain two main parts: theoretical research and practical testing. The theoretical part of this thesis will investigate relevant literature, current approaches, capabilities and limitations of AI systems in understanding and retaining conversational context. In the practical part single and multi-turn testing dialogues will be developed. After collecting the test results evaluation metrics such as Bilingual Evaluation Understudy (BLEU) scores and Dialogue Success Rate (DSR) will be calculated to objectively assess each system's performance. Based on the information received from both parts of this thesis, obtained results will be interpreted and discussed. At the final part, conclusion will be made accordingly.

Keywords

AI, Dialogue systems, Natural language processing, Deep learning, Contextual responses, Machine learning, Chat-bot, Personal assistant, User computer interaction

Jazyk práce: angličtina

Datum vytvoření / odevzdání či podání práce: 17. 3. 2024

Obhajoba závěrečné práce

• Vedoucí: Ing. Jan Pavlík, Ph.D.
• Oponent: Jan Masner, Ph.D.

Česká zemědělská univerzita v Praze

Provozně ekonomická fakulta

Bakalářský studijní program:
Informatics