Year: 2025
Type: conference-paper
Abstract: Large language models (LLMs) have quickly changed the field of artificial intelligence, moving from simple text generation to powering intelligent and autonomous systems. This paper analyzes how LLMs can bring autonomy to modern applications. It begins with an overview of how LLMs work and how they have developed over time. We then compare two main ways to adapt LLMs to specific areas of knowledge: fine-tuning and retrieval-augmented generation (RAG). While fine-tuning updates the model using new data, RAG connects it to external information sources, making it more accurate and easier to update. Next, the paper explores agentic AI, systems where LLMs act as the “brains” of agents that can plan, make decisions, use tools, and work together to reach shared goals. Finally, we discuss current challenges, including coordination, performance, and safety. Overall, the paper shows how LLMs are evolving from chat-based models into the core of intelligent, autonomous systems that can be utilized in real-world environments.
Authors: Matija Špeletić, Stevica Cvetković
Institution: University of Nis
Language: English
Topic: Topic Modeling
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: gold
OpenAlex ID: W7117486691
