🤖 AI Summary
This study addresses the challenges of applying conventional machine learning to physics-dominated manufacturing processes, where experimental data are scarce, expensive, and highly material-specific. The authors propose an integrated framework that combines physical knowledge with data-driven methods to systematically investigate, under limited-data conditions, the impact of data cleaning, feature selection, physics-informed fusion mechanisms, and evaluation strategies on model performance. They innovatively interpret statistically driven feature selection as an explicit modeling assumption and uncover significant instability in model evaluation under small-sample regimes. Experimental results demonstrate that a Gaussian process augmented with residual learning consistently outperforms other approaches, achieving well-calibrated predictive uncertainty—exhibiting an empirical 86% coverage for a nominal 90% prediction interval. However, while residual learning enhances the stability of Gaussian processes, it adversely affects tree-based models.
📝 Abstract
In physically dominated machining processes, experimental datasets are small, expensive, and material-specific; in this regime, data curation, evaluation design, and the form of physics integration can matter as much as the learning algorithm. Using an abrasive waterjet milling dataset ($n{=}155$, Inconel\,718), we make three methodological contributions. First, we separate physics-based data \emph{cleaning} from statistical \emph{curation} and treat the latter as competing modelling hypotheses rather than silent preprocessing. Second, we find that model rankings from a 15-point hold-out set can be unstable: the single-split winner drops from rank~1 to rank~7 under 10-fold cross-validation, while Gaussian Process (GP) variants occupy the top ranks. Third, we study a spectrum of physics integration levels and find that residual learning on a compact physics baseline is competitive for GP, yielding lower variance and an interpretable decomposition, but degrades tree-based models. Bayesian hyper parameter tuning improves parameter-sensitive baselines such as gradient boosting and SVR, yet harms multi-stage hybrid pipelines at this sample size. GP uncertainty intervals are approximately calibrated ($86\%$ empirical coverage at nominal $90\%$). The resulting picture is methodological: for small, expensive process datasets, our results suggest that, in this setting, reliable model comparison benefits from explicit curation hypotheses, robust evaluation, and careful choices about how physics enters the model.