El Agente Quntur: Agentic Active-Space Selection for CASSCF
What it is
This paper introduces a multi-agent system that autonomously selects and refines active spaces for CASSCF calculations in quantum chemistry. It uses LLMs to plan calculations, detect missing orbitals, and revise state assignments, achieving a 48% reduction in mean absolute error compared to deterministic protocols.
The problem it solves
Multireference calculations require substantial human expertise and time to choose and refine active spaces, including selecting orbitals, verifying state descriptions, and ensuring convergence.
What you could use it for
- Automating active space selection for multireference electronic structure calculations
- Refining CASSCF calculations for singlet excitations
- Reducing human expertise requirements in choosing orbital spaces for excited states
Why it matters
Significantly reduces the time and expertise required for multireference calculations, making advanced quantum chemistry methods more accessible and reliable for complex chemical processes.
Read the source ↗Back to the library →
Summary written by an AI model from the source; check the source before you rely on it.