Benchmarking bandgap prediction in semiconductors under experimental and realistic evaluation settings

TL;DR

This paper introduces RealMat-BaG, a benchmark dataset and evaluation framework for assessing the reliability of machine learning models in predicting semiconductor bandgaps under experimental conditions.

Contribution

It provides an open-access experimental bandgap dataset, compares various models, and evaluates their generalization and interpretability in realistic settings.

Findings

Current models have fundamental generalization limitations.

Benchmark reveals poor transfer from DFT to experimental bandgaps.

Analysis highlights interpretability at elemental and structural levels.

Abstract

Accurate bandgap prediction is crucial for semiconductor applications, yet machine learning models trained on computational data often struggle to generalize to experimental bandgap measurements. Challenges related to data fidelity, domain generalization, and model interpretability remain insufficiently addressed in existing evaluation frameworks. To bridge this gap, we introduce RealMat-BaG, a benchmark for assessing model reliability under experimentally relevant conditions. We curate an open-access dataset of experimental bandgaps with aligned crystal structures and compare graph neural networks as well as classical machine learning baselines. Our framework evaluates performance across statistical and domain-based splits, examines transfer from DFT-computed to experimental bandgaps, and analyzes interpretability at both elemental-property and structural levels. Our results reveal the fundamental generalization limitations of current bandgap prediction models and establish a benchmark aligned with experimental measurements for developing more reliable learning strategies for materials discovery.