Evaluating Machine Learning Models for Molecular Property Prediction: Performance and Robustness on Out-of-Distribution Data

Poster presentation at the 13th International Conference on Chemical Structures (ICCS) summarizing our systematic evaluation of 14 machine learning models across eight datasets and ten splitting strategies, examining how the choice of OOD generation procedure shapes both absolute performance and the strength of the ID–OOD correlation. See the associated publication.

Authors: Hosein Fooladi, Thi Ngoc Lan Vu, and Johannes Kirchmair.