🤖 AI Summary
Constructing machine learning interatomic potentials (MLIPs) that simultaneously achieve high accuracy, dynamical stability, and computational efficiency entails balancing multiple objectives, which cannot be adequately captured by a single loss function. This work proposes MLIPilot, a framework that leverages large language models—such as GPT-4.1 and Qwen3-32B—as autonomous research agents to automate MLIP development. Integrated with a physics-informed constraint scoring card and the Atomic Simulation Environment (ASE) toolchain, the agent autonomously formulates hypotheses, modifies training code, and submits computational jobs on MACE potentials using the QM7 and Cu EMT datasets. The approach enhances both automation and auditability in MLIP development, substantially reducing manual trial-and-error. The most capable agent successfully transformed an initially non-compliant model into one satisfying all physical constraints, uncovering effective strategies including output normalization, loss function tuning, progressive training schedules, and model capacity optimization.
📝 Abstract
Constructing production-quality machine-learned interatomic potentials (MLIPs) requires balancing accuracy, dynamical stability, and computational throughput under constraints that are not captured by a single training loss. We introduce MLIPilot, an auto-research framework in which tool-calling large language models propose hypotheses, edit MLIP training code, launch HPC jobs, and accept or revert changes using a fixed, physically constrained scorecard. We evaluate MLIPilot on MACE potential optimization using both commercial and open-weight LLM agents, including GPT-5.5, GPT-4.1, Mistral-24B, and Qwen3-32B. The benchmarks span molecular and periodic settings: a QM7-derived dataset for which we generated B3LYP/6-31G(d) energies and forces, and a Cu EMT dataset with periodic copper supercells labeled by ASE's Effective Medium Theory calculator. Across these benchmarks, the strongest agents move initially constraint-violating baselines to accepted models by discovering useful training strategies, including output normalization, loss-function changes, progressive training schedules, and model-capacity adjustments. These results suggest that LLM agents can serve as autonomous operators for scientific machine-learning workflows when their search is constrained by domain-specific validation criteria, shifting part of MLIP development from manual trial-and-error toward auditable, automated experimentation.