5 papers
SCORE: Replacing Layer Stacking with Contractive Recurrent Depth
Guillaume Godin
Residual connections are central to modern deep neural networks, enabling stable optimization and efficient information flow across depth. In this work, we propose SCORE (Skip-Conn…
Fast Leave-One-Out Approximation from Fragment-Target Prevalence Vectors (molFTP) : From Dummy Masking to Key-LOO for Leakage-Free Feature Construction
Guillaume Godin
We introduce molFTP (molecular fragment-target prevalence), a compact representation that delivers strong predictive performance. To prevent feature leakage across cross-validation…
Bond-Centered Molecular Fingerprint Derivatives: A BBBP Dataset Study
Guillaume Godin
Bond Centered FingerPrint (BCFP) are a complementary, bond-centric alternative to Extended-Connectivity Fingerprints (ECFP). We introduce a static BCFP that mirrors the bond-convol…
Z-Error Loss for Training Neural Networks
Guillaume Godin
Outliers introduce significant training challenges in neural networks by propagating erroneous gradients, which can degrade model performance and generalization. We propose the Z-E…
All You Need Is Synthetic Task Augmentation
Guillaume Godin
Injecting rule-based models like Random Forests into differentiable neural network frameworks remains an open challenge in machine learning. Recent advancements have demonstrated t…