MoleculeNet: a benchmark for molecular machine learning
By Zhenqin Wu, Bharath Ramsundar, Evan N. Feinberg, Joseph Gomes, Caleb Geniesse, Aneesh S. Pappu, K. Leswing, Vijay S. Pande. 2017
2,566 citations recorded by the source corpus.
A large scale benchmark for molecular machine learning consisting of multiple public datasets, metrics, featurizations and learning algorithms.
Subfields: Computational Theory and Mathematics, Materials Chemistry, Molecular Biology, Evaluation & Benchmarks, Computational Chemistry & Drug Discovery
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Computational Theory and Mathematics
- Tommi S. Jaakkola
- Gisbert Schneider
- Regina Barzilay
- Francesca Grisoni
- Tess E. Smidt
- Charles Siegel
- Abhinav Vishnu
- Garrett B. Goh
In Materials Chemistry
- Tommi S. Jaakkola
- Regina Barzilay
- Gisbert Schneider
- Jure Leskovec
- Tess E. Smidt
- Francesca Grisoni
- Jian Tang
- Wengong Jin
In Molecular Biology
- Tommi S. Jaakkola
- Regina Barzilay
- Frank No'e
- Tess E. Smidt
- Stephan Günnemann
- Charles Siegel
- Jure Leskovec
- Chaopeng Shen
In Evaluation & Benchmarks
- Sergey Levine
- Chelsea Finn
- Kyle Swanson
- Charles Siegel
- Abhinav Vishnu
- Garrett B. Goh
- Nathan Oken Hodas
- Percy Liang
In Computational Chemistry & Drug Discovery
- Tess E. Smidt
- Tommi S. Jaakkola
- Chaopeng Shen
- Mario Geiger
- Regina Barzilay
- Gisbert Schneider
- Francesca Grisoni
- Stephan Günnemann
Citations in the corpus
- Strategies for Pre-training Graph Neural Networks (2019, methodology): “On the HIV dataset, where a number of recent works (Wu et al., 2018; Li et al., 2017; Ishiguro et al., 2019) have reported performance on the same scaffold split and using the same protocol, our best pre-trained model (ContextPred + Supervised) achieves state-of-the-art performance.”
- Directional Message Passing for Molecular Graphs (2020, background): “…a wide variety of molecular properties, both low-level quantum mechanical properties such as potential energy, energy of the highest occupied molecular orbital (HOMO), and the dipole moment and high-level properties such as toxicity, permeability, and adverse drug reactions (Wu et al., 2018).”
- Mol2vec: Unsupervised Machine Learning Approach with Chemical Intuition (2018, positive): “Mol2vec yields better predictions (R 2ext = 0.86) than the originally reported multiple linear regression (MLR) model 25 as well as a molecular graph convolution method, 36 and importantly outperforms our Morgan FP based baseline model (R 2ext = 0.66).”
- Learning Graph-Level Representation for Drug Discovery (2017, methodology): “For most macroscopic molecular properties predication tasks, like efficacy and toxicity, we could neglect the detailed edge information and take a molecule as an undirected graph(Wu et al. 2017).”
- Opportunities and obstacles for deep learning in biology and medicine (2017, methodology): “Broader recent benchmarking has shown that relative merits of these methods depends on the dataset and cross validation strategy [312], though evaluation often uses auROC (area under the receiver operating characteristic curve), which has limited utility due to the large active/inactive class imbala”
- Using Rule-Based Labels for Weak Supervised Learning: A ChemNet for Transferable Chemical Property Prediction (2017, methodology): “To ensure that our results are comparable with contemporary DNN models reported in the literature [34] and earlier work on Chemception CNN models [15, 16] and SMILES2vec RNN models [13], we used the Tox21, HIV, and FreeSolv dataset from the MoleculeNet benchmark [34] for predicting toxicity,…”
- Relevance of Rotationally Equivariant Convolutions for Predicting Molecular Properties (2020, background): “1 Introduction The discovery of novel molecules has been accelerated by advances in computational quantum chemistry and machine learning assisted exploration of chemical space [1, 2, 3].”
- Generalizing Convolutional Neural Networks for Equivariance to Lie Groups on Arbitrary Continuous Data (2020, methodology): “Now we apply LieConv to the QM9 molecular property learning task (Wu et al., 2018).”
- Set2Graph: Learning Graphs From Sets (2020, methodology): “We used graph datasets from (Wu et al., 2018) us-ing the Open Graph Benchmark (OGB, 2019).”
- Provably Powerful Graph Networks (2019, methodology): “For the regression task, we conducted an experiment on a standard graph learning benchmark called the QM9 dataset (Ramakrishnan et al., 2014; Wu et al., 2018).”
- WILDS: A Benchmark of in-the-Wild Distribution Shifts (2020, positive): “While OGB-MolPCBA primarily focuses on predicting biophysical activity (e.g., protein binding), other datasets in MoleculeNet (Wu et al., 2018) include prediction of quantum mechanical properties (e.g., HOMO/LUMO), physical chemistry properties (e.g., water solubility), and physiological properties…”
- Automated Machine Learning: From Principles to Practices (2018, background): “In this application, experience ( 𝐸 ) corresponds to molecules with its property (label) and 𝑃 measures the performance of molecular property prediction, such as accuracy or AUC [150].”
- Recent advances and applications of deep learning methods in materials science (2021, neutral): “com/deepchem/deepchem [107] dgl-lifesci Prot https://github.”