# Most Ligand-Based Classification Benchmarks Reward Memorization Rather than Generalization

arXiv: https://arxiv.org/abs/1706.06619 · HTML: https://mlsanity.com/paper/1706.06619/

## Key facts

- Cited by 5 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 20 Jun 2017.

## Built on by (most cited first, 5 of 5)

- [Quantum Machine Learning Framework for Virtual Screening in Drug Discovery: a Prospective Quantum Advantage](https://mlsanity.com/paper/2204.04017/) (2022)
- [Assigning Confidence to Molecular Property Prediction](https://mlsanity.com/paper/2102.11439/) (2021)
- [Machine Learning Small Molecule Properties in Drug Discovery](https://mlsanity.com/paper/2308.12354/) (2023)
- [HD-Bind: Encoding of Molecular Structure with Low Precision, Hyperdimensional Binary Representations](https://mlsanity.com/paper/2303.15604/) (2023)
- [ALMERIA: Boosting pairwise molecular contrasts with scalable methods](https://mlsanity.com/paper/2305.13254/) (2023)

## Abstract (arXiv)

Undetected overfitting can occur when there are significant redundancies between training and validation data. We describe AVE, a new measure of training-validation redundancy for ligand-based classification problems that accounts for the similarity amongst inactive molecules as well as active. We investigated seven widely-used benchmarks for virtual screening and classification, and show that the amount of AVE bias strongly correlates with the performance of ligand-based predictive methods irrespective of the predicted property, chemical fingerprint, similarity measure, or previously-applied unbiasing techniques. Therefore, it may be that the previously-reported performance of most ligand-based methods can be explained by overfitting to benchmarks rather than good prospective accuracy.