AI Agent with Model Selection, Tool Synthesis, and Proactive Knowledge Base Management
This paper presents a concept and pilot implementation of an adaptive AI agent that extends conventional large language model (LLM) systems with planning, tool usage, long-term memory, and continuous self-improvement. Unlike static agent architectures, the proposed solution operates on three levels of adaptability: ope...