Autonomous AI for Material Discovery
Scientific discovery is increasingly moving toward autonomous research systems that integrate artificial intelligence with automated experimentation and computation. In our group, we explore the development of autonomous AI systems capable of guiding the discovery and optimization of molecules and materials with minimal human intervention.
Our research focuses on building AI agents and decision-making frameworks that can plan experiments, evaluate results, and iteratively improve candidate materials. These systems integrate machine learning models, optimization algorithms, and scientific reasoning to create intelligent discovery pipelines. By connecting AI models with computational simulations and experimental platforms, we aim to enable closed-loop materials discovery systems that accelerate research cycles and significantly reduce the time required to discover new functional materials.
Applying autonomous AI systems to accelerate chemical research, including reaction prediction, materials discovery, and scientific decision-making.
Developing agent-based AI frameworks that can plan, reason, and coordinate scientific workflows for materials discovery.
Connecting AI agents with automated experimental platforms and self-driving laboratories (SDL) to enable closed-loop scientific discovery.
Selected papers
[1] Y. Kang and J. Kim*
Nature Communications, 2024
[2] J. Bai#, A. Aldossary#, T. Swanick, M. Müller, Y. Kang, J. Zhang, J. W. Lee, T. W. Ko, M. G. Vakili, V. Bernales*, A. Aspuru-Guzik*
El Agente Gráfico: Structured Execution Graphs for Scientific Agents
Arxiv, 2026x
[3] C. Choi#, Y. Zou#, M. Müller, H. Hao, Y. Kang, J. B. Pérez-Sánchez, I. Gustin, H. Xu, M. G. Vakili, C. Crebolder, A. Aspuru-Guzik*, V. Bernales*
El Agente Estructural: An Artificially Intelligent Molecular Editor
Arxiv, 2026