Research Overview

AiX Lab – Research Mission & Vision

AiX Lab at Gyeongsang National University advances trustworthy AI through mathematically grounded verification, safe domain-specialized LLMs, clinically meaningful medical intelligence, and real-world autonomous racing platforms.

Our Mission

Develop AI systems that are safe, interpretable, and dependable in real-world conditions.
Build domain-specialized LLMs that achieve trustworthy reasoning and reduced hallucination.
Create clinically meaningful Medical AI for ophthalmology and healthcare decision support.
Advance autonomous robot racing and safe high-speed navigation under uncertainty.

Our Vision

We strive to build AI technologies that are:

Provably safe through formal verification and runtime assurance.
Interpretable and aligned for transparent and accountable decision-making.
Clinically useful through noise-robust medical modeling and domain LLMs.
Fast and reliable in high-speed autonomous robotics.
Educationally impactful by training the next generation of AI and CPS engineers.

Our Key Strengths

Formal Methods for AI Safety
Timed automata, process algebra, and model checking for provably safe and correct AI-driven systems, including verified code generation and runtime monitoring.
Safe & Domain-Specialized LLMs
Compact and trustworthy LLMs (Ophtimus Series) designed for interpretability, alignment, and reduced hallucination with applications in medicine and engineering.
Medical AI & Oculomics
Deep learning and LLM-augmented models for ophthalmology, oculomics, clinical decision support, and noise-robust diagnostic modeling.
Robot Racing & Autonomous Driving
Research and education using F1TENTH and RoboRacer platforms focusing on safe high-speed navigation, SLAM, motion planning, reinforcement learning, and real-world uncertainty handling.