mlsanityData to 14 Oct 2026

Fast Leave-One-Out Approximation from Fragment-Target Prevalence Vectors (molFTP) : From Dummy Masking to Key-LOO for Leakage-Free Feature Construction

arXiv 2510.06029 ↗First posted 7 Oct 2025cs.LG cs.AIcited by 0

Not on the map. Why?

Abstract (arXiv)

We introduce molFTP (molecular fragment-target prevalence), a compact representation that delivers strong predictive performance. To prevent feature leakage across cross-validation folds, we implement a dummy-masking procedure that removes information about fragments present in the held-out molecules. We further show that key leave-one-out (key-loo) closely approximates true molecule-level leave-one-out (LOO), with deviation below 8% on our datasets. This enables near full data training while preserving unbiased cross-validation estimates of model performance. Overall, molFTP provides a fast, leakage-resistant fragment-target prevalence vectorization with practical safeguards (dummy masking or key-LOO) that approximate LOO at a fraction of its cost.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 13 of its 40 references that are in the map. First posted on arXiv 7 Oct 2025.

Builds on

13 of 40 references.

References · 13

  1. 2020A Simple Framework for Contrastive Learning of Visual Representationscited by 12,049
  2. 2016XGBoost: A Scalable Tree Boosting Systemcited by 5,166
  3. 2025Why Language Models Hallucinatecited by 220
  4. 2024On the Scalability of GNNs for Molecular Graphscited by 36
  5. 2024Beyond Random Missingness: Clinically Rethinking for Healthcare Time Series Imputationcited by 6
  6. 2024Is Training Data Quality or Quantity More Impactful to Small Language Model Performance?cited by 4
  7. 2023Approximate Leave-one-out Cross Validation for Regression with ℓ1 Regularizers (extended version)cited by 3
  8. 2024Sort & Slice: A Simple and Superior Alternative to Hash-Based Folding for Extended-Connectivity Fingerprintscited by 3
  9. 2020Binary classification with ambiguous training datacited by 1
  10. 2023Improving classifier decision boundaries using nearest neighborscited by 1
  11. 2025Simultaneous analysis of approximate leave-one-out cross-validation and mean-field inferencecited by 1
  12. 2025Z-Error Loss for Training Neural Networkscited by 1
1 more
  1. 2025Bond-Centered Molecular Fingerprint Derivatives: A BBBP Dataset Studycited by 1

Built on by

Not cited yet.