
KM
Kareem M. Gameel, Ihor Neporozhnii, Sjoerd Hoogland, Oleksandr Voznyy
· 1 min read
ResearcharXiv cs.LG
The Mechanics of Delta Learning: Target Design for Generalizable Scientific Machine Learning
arXiv:2609.28782v1 Announce Type: new
Abstract: In scientific machine learning, $\Delta$-learning trains models on residual errors relative to physical baselines, assuming that more accurate baselines with smaller residual scales inherently improve downstream performance. Here, we demonstrate that residual scale alone is an insufficient heuristic for learnability. Evaluating molecular graph neural networks on total energy targets, we show that complex local descriptor baselines can yield small residual targets that are disproportionately rough within architecture-informed proxy spaces and harder to learn relative to their scale. Conversely, semi-empirical baseline reduces both scale and normalized roughness, improving in-domain and out-of-domain prediction. We introduce scale-normalized graph Dirichlet roughness ($D_{\text{IQR}}$) as a pre-training diagnostic for residual learnability and establish baseline complementarity as a core target-design principle, elevating target space formulation alongside model architecture as a key axis for scientific machine learning.
Original source
This story was published by arXiv cs.LG and written by Kareem M. Gameel, Ihor Neporozhnii, Sjoerd Hoogland, Oleksandr Voznyy. SyncAI.news shows a preview; the complete article is on the publisher's site.
Read the full story on arxiv.org


