DREAMS: Density Functional Theory Based Research Engine for Agentic Materials Simulation

2025-07-18

AI Summary

Material discovery has long been hindered by the heavy reliance of density functional theory (DFT) simulations on expert knowledge—particularly for parameter tuning, systematic error correction, and convergence criterion assessment. To address this, we propose the first hierarchical multi-agent framework tailored for high-throughput, high-fidelity materials simulation. It integrates large language models (LLMs) for task planning and contextual coherence, while coordinating domain-specific agents that perform DFT convergence testing, HPC job scheduling, systematic error modeling, and Bayesian ensemble sampling. The framework achieves Level-3 autonomous exploration capability and incorporates a hallucination-mitigating shared canvas mechanism. Evaluated on the Sol27LC benchmark, it attains prediction errors below 1%, successfully resolves the CO/Pt(111) adsorption configuration, and—uniquely—quantifies functional-choice-induced uncertainty with expert-level accuracy.

Technology Category

• Search and Optimization: Sampling/Simulation-based Search
• Multiagent Systems: Multiagent Planning
• Cognitive Modeling & Cognitive Systems: Agent Architectures

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Search and Retrieval-Augmented AI: Agentic search
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems

Abstract

Materials discovery relies on high-throughput, high-fidelity simulation techniques such as Density Functional Theory (DFT), which require years of training, extensive parameter fine-tuning and systematic error handling. To address these challenges, we introduce the DFT-based Research Engine for Agentic Materials Screening (DREAMS), a hierarchical, multi-agent framework for DFT simulation that combines a central Large Language Model (LLM) planner agent with domain-specific LLM agents for atomistic structure generation, systematic DFT convergence testing, High-Performance Computing (HPC) scheduling, and error handling. In addition, a shared canvas helps the LLM agents to structure their discussions, preserve context and prevent hallucination. We validate DREAMS capabilities on the Sol27LC lattice-constant benchmark, achieving average errors below 1% compared to the results of human DFT experts. Furthermore, we apply DREAMS to the long-standing CO/Pt(111) adsorption puzzle, demonstrating its long-term and complex problem-solving capabilities. The framework again reproduces expert-level literature adsorption-energy differences. Finally, DREAMS is employed to quantify functional-driven uncertainties with Bayesian ensemble sampling, confirming the Face Centered Cubic (FCC)-site preference at the Generalized Gradient Approximation (GGA) DFT level. In conclusion, DREAMS approaches L3-level automation - autonomous exploration of a defined design space - and significantly reduces the reliance on human expertise and intervention, offering a scalable path toward democratized, high-throughput, high-fidelity computational materials discovery.

Problem

• Automates complex DFT simulations to reduce human expertise dependency
• Solves materials discovery challenges with multi-agent LLM framework
• Validates accuracy on benchmarks like Sol27LC and CO/Pt(111)

Innovation

• Hierarchical multi-agent framework with LLM planner
• Shared canvas for agent context and hallucination prevention
• Automated DFT simulation with minimal human intervention