# Migration as a Probe: A Generalizable Benchmark Framework for Specialist vs. Generalist Machine-Learned Force Fields

## TL;DR

This paper introduces a benchmarking framework using defect migration pathways to evaluate machine-learned force fields, comparing specialist, generalist, and hybrid models in materials science.

## Contribution

It presents a novel migration-based benchmarking framework that assesses different training paradigms of MLFFs, providing insights into their interpolation and extrapolation capabilities.

## Key findings

- Fine-tuned models outperform from-scratch models in kinetic tasks.
- Different training strategies learn distinct physical aspects.
- Migration-based probes effectively diagnose model performance and representations.

## Abstract

Machine-learned force fields (MLFFs), especially pre-trained foundation models, are transforming computational materials science by enabling ab initio-level accuracy at molecular dynamics scales. Yet their rapid rise raises a key question: should researchers train specialist models from scratch, fine-tune generalist foundation models, or use hybrid approaches? The trade-offs in data efficiency, accuracy, cost, and robustness to out-of-distribution failure remain unclear. We introduce a benchmarking framework using defect migration pathways, evaluated through nudged elastic band trajectories, as diagnostic probes that test both interpolation and extrapolation. Using Cr-doped Sb2Te3 as a representative two-dimensional material, we benchmark multiple training paradigms within the MACE architecture across equilibrium, kinetic (atomic migration), and mechanical (interlayer sliding) tasks. Fine-tuned models substantially outperform from-scratch and zero-shot approaches for kinetic properties but show partial loss of long-range physics. Representational analysis reveals distinct, non-overlapping latent encodings, indicating that different training strategies learn different aspects of system physics. This framework provides practical guidelines for MLFF development and establishes migration-based probes as efficient diagnostics linking performance to learned representations, guiding future uncertainty-aware active learning.