Skip to content
Branch: master
Find file History
mufeili [Model Zoo] AttentiveFP (#955)
* Update

* Fix style

* Update

* Update

* Fix

* Update

* Update
Latest commit 52c7ef4 Oct 25, 2019
Type Name Latest commit message Commit time
Failed to load latest commit information. [Model Zoo] AttentiveFP (#955) Oct 25, 2019 [Model Zoo] AttentiveFP (#955) Oct 25, 2019 [Model Zoo] AttentiveFP (#955) Oct 25, 2019 [Model Zoo] AttentiveFP (#955) Oct 25, 2019 [Model Zoo] AttentiveFP (#955) Oct 25, 2019

Property Prediction


Classification tasks require assigning discrete labels to a molecule, e.g. molecule toxicity.


  • Tox21. The "Toxicology in the 21st Century" (Tox21) initiative created a public database measuring toxicity of compounds, which has been used in the 2014 Tox21 Data Challenge. The dataset contains qualitative toxicity measurements for 8014 compounds on 12 different targets, including nuclear receptors and stress response pathways. Each target yields a binary prediction problem. MoleculeNet [1] randomly splits the dataset into training, validation and test set with a 80/10/10 ratio. By default we follow their split method.


  • Graph Convolutional Network [2], [3]. Graph Convolutional Networks (GCN) have been one of the most popular graph neural networks and they can be easily extended for graph level prediction. MoleculeNet [1] reports baseline results of graph convolutions over multiple datasets.
  • Graph Attention Networks [7]. Graph Attention Networks (GATs) incorporate multi-head attention into GCNs, explicitly modeling the interactions between adjacent atoms.


Use with arguments

-m {GCN, GAT}, MODEL, model to use
-d {Tox21}, DATASET, dataset to use

If you want to use the pre-trained model, simply add -p.

We use GPU whenever it is available.


GCN on Tox21

Source Averaged Test ROC-AUC Score
MoleculeNet [1] 0.829
DeepChem example 0.813
Pretrained model 0.826

Note that the dataset is randomly split so these numbers are only for reference and they do not necessarily suggest a real difference.

GAT on Tox21

Source Averaged Test ROC-AUC Score
Pretrained model 0.827


Regression tasks require assigning continuous labels to a molecule, e.g. molecular energy.


  • Alchemy. The Alchemy Dataset is introduced by Tencent Quantum Lab to facilitate the development of new machine learning models useful for chemistry and materials science. The dataset lists 12 quantum mechanical properties of 130,000+ organic molecules comprising up to 12 heavy atoms (C, N, O, S, F and Cl), sampled from the GDBMedChem database. These properties have been calculated using the open-source computational chemistry program Python-based Simulation of Chemistry Framework (PySCF). The Alchemy dataset expands on the volume and diversity of existing molecular datasets such as QM9.
  • PubChem BioAssay Aromaticity. The dataset is introduced in Pushing the Boundaries of Molecular Representation for Drug Discovery with the Graph Attention Mechanism, for the task of predicting the number of aromatic atoms in molecules. The dataset was constructed by sampling 3945 molecules with 0-40 aromatic atoms from the PubChem BioAssay dataset.


  • Message Passing Neural Network [6]. Message Passing Neural Networks (MPNNs) have reached the best performance on the QM9 dataset for some time.
  • SchNet [4]. SchNet employs continuous filter convolutional layers to model quantum interactions in molecules without requiring them to lie on grids.
  • Multilevel Graph Convolutional Neural Network [5]. Multilevel Graph Convolutional Neural Networks (MGCN) are hierarchical graph neural networks that extract features from the conformation and spatial information followed by the multilevel interactions.
  • AttentiveFP [8]. AttentiveFP combines attention and GRU for better model capacity and shows competitive performance across datasetts.


Use with arguments

-m {MPNN, SCHNET, MGCN, AttentiveFP}, Model to use
-d {Alchemy, Aromaticity}, Dataset to use

If you want to use the pre-trained model, simply add -p. Currently we only support pre-trained models of AttentiveFP on PubChem BioAssay Aromaticity dataset.



The Alchemy contest is still ongoing. Before the test set is fully released, we only include the performance numbers on the training and validation set for reference.

Model Training MAE Validation MAE
SchNet [4] 0.2665 0.6139
MGCN [5] 0.2395 0.6463
MPNN [6] 0.2452 0.6259

PubChem BioAssay Aromaticity

Model Test RMSE
AttentiveFP [8] 0.6998


[8] visualizes the weights of atoms in readout for possible interpretations like the figure below. We provide a jupyter notebook for performing the visualization and you can download it with wget

Dataset Customization

To customize your own dataset, see the instructions here.


[1] Wu et al. (2017) MoleculeNet: a benchmark for molecular machine learning. Chemical Science 9, 513-530.

[2] Duvenaud et al. (2015) Convolutional networks on graphs for learning molecular fingerprints. Advances in neural information processing systems (NeurIPS), 2224-2232.

[3] Kipf et al. (2017) Semi-Supervised Classification with Graph Convolutional Networks. The International Conference on Learning Representations (ICLR).

[4] Schütt et al. (2017) SchNet: A continuous-filter convolutional neural network for modeling quantum interactions. Advances in Neural Information Processing Systems (NeurIPS), 992-1002.

[5] Lu et al. (2019) Molecular Property Prediction: A Multilevel Quantum Interactions Modeling Perspective. The 33rd AAAI Conference on Artificial Intelligence.

[6] Gilmer et al. (2017) Neural Message Passing for Quantum Chemistry. Proceedings of the 34th International Conference on Machine Learning, JMLR. 1263-1272.

[7] Veličković et al. (2018) Graph Attention Networks. The International Conference on Learning Representations (ICLR).

[8] Xiong et al. (2019) Pushing the Boundaries of Molecular Representation for Drug Discovery with the Graph Attention Mechanism. Journal of Medicinal Chemistry.

You can’t perform that action at this time.