Joint NJIT Physics & Rutgers Physics Seminar

October 23rd, Friday (*SPECIAL DAY*)

Automating Strongly Correlated Simulations with AI Agents

Prof. Chong Sun

Rutgers University–New Brunswick

(Materials Phys., Host: Rutgers-Newark Physics Dept.)

SPECIAL ROOM at RUTGERS NEWARK CAMPUS:

Dana Room (4th floor of Dana Library) map

(**BRING YOUR PHOTO ID TO ENTER BUILDING**)

SPECIAL TIME: 11:30 am - 12:30 pm (Coffee: 11am)

The development of artificial intelligence (AI) is helping bridge interdisciplinary knowledge gaps. Agentic frameworks that combine large language models (LLMs) with expert-defined skills are making specialized tasks more accessible by automating workflows while providing reasoning and guidance. Computational science has been an early adopter of this technology, with numerous frameworks proposed over the past year for automating workflows in quantum chemistry and other domains. However, strongly correlated electronic-structure problems, as a major challenge in computational science, remain largely unexplored. In this talk, we introduce CAFES, a framework for strongly correlated computational studies. CAFES addresses two key challenges: the fragmentation of many-body software and the difficulty of selecting appropriate methods across different correlation regimes. It integrates correlation diagnostics, adaptive method selection, and LLM assistance for molecular, crystalline, and model-Hamiltonian systems. Its study–task organization separates high-level planning from task-level execution, while a curated scientific knowledge layer provides reusable guidance for method selection, workflow design, and result interpretation. We demonstrate CAFES through research-level applications, including active space DMRG simulations of excited states in a large molecule and phase exploration of the honeycomb Hubbard model with DMET simulations.