TSI logo 25 years
LV EN
Home > About Us > Structure and Government > Abstracts

DEGREE

PROGRAMME

FACULTY

YEAR

LANGUAGE

KEYWORDS

Development of a Presentation Generation Web Service Using AI Language Models

This bachelor's thesis presents the development of a presentation generation web service using AI language models. The application integrates APIs such as OpenAI for text generation, Pexels for image retrieval, and Google Slides for presentation creation, providing a comprehensive tool for generating high-quality presentations and videos. The backend is built using Flask, and the frontend uses React, ensuring a seamless user experience. Key features include user authentication, state management, and dynamic content generation.The project involved analyzing existing AI-powered presentation tools, developing the web service with frontend and backend technologies, and integrating various APIs to enhance functionality. Rigorous testing ensured stability, efficiency, and user-friendliness. The resulting application can generate high-quality presentations and convert them into engaging videos with minimal effort.This thesis demonstrates AI's potential in enhancing digital content creation, offering significant improvements in efficiency and user engagement. The developed web service provides a valuable tool for users needing to quickly create professional presentations.

Author: Igors Pticins

Supervisor: Aleksejs Vesjolijs

Degree: Bachelor

Year: 2024

Work Language: English

Study programme: Computer Science

More...

Enhancement Strategies for Retrieval-Augmented Generation Systems

This thesis systematically explores the enhancement of Retrieval-Augmented Generation (RAG) systems within Large Language Models, emphasizing optimization of retrieval parameters and generation accuracy. We investigate optimal configurations in RAG systems, including chunk size and overlap percentages, top-k selection, query transformations, different retrieval methods, different LLMs, namely GPT-3.5-Turbo and GPT-4, discovering that a chunk size of 500 tokens generally offers the best performance. Vector search using cosine similarity emerges as the most effective retrieval method, significantly enhancing both context precision and recall across various tasks and knowledge bases. Experimentation within the CRUD-RAG framework demonstrates its applicability in diverse tasks from content creation to knowledge refinement. Our findings indicate that enhancements in retrieval settings can markedly improve the performance of RAG systems, making them more efficient and adaptable for complex information synthesis and retrieval tasks. These results affirm the potential of systematic enhancements to improve AI-driven language models in practical applications, contributing significant insights and practical approaches to the evolving landscape of RAG system research.

Author: Sigita Lapiņa

Supervisor: Dmitry Pavlyuk

Degree: Master

Year: 2024

Work Language: English

Study programme: Computer Sciences

More...

Table View
Text View