HoloMind: A Minimal Autonomous Learning System Driven by Curiosity, Confidence, and Self-Termination

HoloMind is a minimal autonomous learning system capable of self-directed learning, curiosity-driven inquiry, and confidence-regulated learning termination. It operates without external reward functions, reinforcement learning, or human feedback. The system architecture includes an embedding-first memory design (STM, MTM, LTM), a curiosity engine, and a reasoning module, demonstrating autonomous learning behavior validated through reproducible test logs.