MolGAN: An implicit generative model for small molecular graphs

eep generative models for graph-structured data offer a new angle on the problem of chemical synthesis: by optimizing differentiable models that directly generate molecular graphs, it is pos-sible to side-step expensive search procedures in the discrete and vast space of chemical structures. We introduce MolGAN, an implicit, likelihood-free generative model for small molecular graphs that circumvents the need for expensive graph matching procedures or node ordering heuris-tics of previous likelihood-based methods. Our method adapts generative adversarial networks (GANs) to operate directly on graph-structured data. We combine our approach with a reinforce-ment learning objective to encourage the genera-tion of molecules with specific desired chemical properties. In experiments on the QM9 chemi-cal database, we demonstrate that our model is capable of generating close to 100% valid com-pounds. MolGAN compares favorably both to recent proposals that use string-based (SMILES) representations of molecules and to a likelihood-based method that directly generates graphs, al-beit being susceptible to mode collapse.

[1]  Wojciech Zaremba,et al.  Improved Techniques for Training GANs , 2016, NIPS.

[2]  Pavlo O. Dral,et al.  Quantum chemistry structures and properties of 134 kilo molecules , 2014, Scientific Data.

[3]  Mike Preuss,et al.  Planning chemical syntheses with deep neural networks and symbolic AI , 2017, Nature.

[4]  R. J. Williams,et al.  Simple Statistical Gradient-Following Algorithms for Connectionist Reinforcement Learning , 2004, Machine Learning.

[5]  Max Welling,et al.  Auto-Encoding Variational Bayes , 2013, ICLR.

[6]  John Comer,et al.  Lipophilicity Profiles: Theory and Measurement , 2007 .

[7]  Léon Bottou,et al.  Wasserstein Generative Adversarial Networks , 2017, ICML.

[8]  Daan Wierstra,et al.  Stochastic Backpropagation and Approximate Inference in Deep Generative Models , 2014, ICML.

[9]  Yibo Li,et al.  Multi-objective de novo drug design with conditional graph generative model , 2018, Journal of Cheminformatics.

[10]  Aaron C. Courville,et al.  Improved Training of Wasserstein GANs , 2017, NIPS.

[11]  Jure Leskovec,et al.  GraphRNN: A Deep Generative Model for Graphs , 2018, ICML 2018.

[12]  Richard S. Zemel,et al.  Gated Graph Sequence Neural Networks , 2015, ICLR.

[13]  Jure Leskovec,et al.  Representation Learning on Graphs: Methods and Applications , 2017, IEEE Data Eng. Bull..

[14]  Daniel D. Johnson,et al.  Learning Graphical State Transitions , 2016, ICLR.

[15]  Max Welling,et al.  Modeling Relational Data with Graph Convolutional Networks , 2017, ESWC.

[16]  Niloy Ganguly,et al.  Designing Random Graph Models Using Variational Autoencoders With Applications to Chemical Design , 2018, ArXiv.

[17]  Jimmy Ba,et al.  Adam: A Method for Stochastic Optimization , 2014, ICLR.

[18]  Nitish Srivastava,et al.  Dropout: a simple way to prevent neural networks from overfitting , 2014, J. Mach. Learn. Res..

[19]  David Weininger,et al.  SMILES, a chemical language and information system. 1. Introduction to methodology and encoding rules , 1988, J. Chem. Inf. Comput. Sci..

[20]  Nicola De Cao,et al.  Hyperspherical Variational Auto-Encoders , 2018, UAI 2018.

[21]  Regina Barzilay,et al.  Junction Tree Variational Autoencoder for Molecular Graph Generation , 2018, ICML.

[22]  Peter Ertl,et al.  Estimation of synthetic accessibility score of drug-like molecules based on molecular complexity and fragment contributions , 2009, J. Cheminformatics.

[23]  Stefano Ermon,et al.  Graphite: Iterative Generative Modeling of Graphs , 2018, ICML.

[24]  Stephan Günnemann,et al.  NetGAN: Generating Graphs via Random Walks , 2018, ICML.

[25]  Razvan Pascanu,et al.  Learning Deep Generative Models of Graphs , 2018, ICLR 2018.

[26]  Gisbert Schneider,et al.  Computer-based de novo design of drug-like molecules , 2005, Nature Reviews Drug Discovery.

[27]  Ben Poole,et al.  Categorical Reparameterization with Gumbel-Softmax , 2016, ICLR.

[28]  Yee Whye Teh,et al.  The Concrete Distribution: A Continuous Relaxation of Discrete Random Variables , 2016, ICLR.

[29]  Minyi Guo,et al.  GraphGAN: Graph Representation Learning with Generative Adversarial Nets , 2017, AAAI.

[30]  G. V. Paolini,et al.  Quantifying the chemical beauty of drugs. , 2012, Nature chemistry.

[31]  Steven Skiena,et al.  Syntax-Directed Variational Autoencoder for Molecule Generation , 2017 .

[32]  Nikos Komodakis,et al.  GraphVAE: Towards Generation of Small Graphs Using Variational Autoencoders , 2018, ICANN.

[33]  Joan Bruna,et al.  Spectral Networks and Locally Connected Networks on Graphs , 2013, ICLR.

[34]  Yoshua Bengio,et al.  Generative Adversarial Nets , 2014, NIPS.

[35]  Guy Lever,et al.  Deterministic Policy Gradient Algorithms , 2014, ICML.

[36]  Mike Preuss,et al.  Learning to Plan Chemical Syntheses , 2017, ArXiv.

[37]  Pierre Vandergheynst,et al.  Geometric Deep Learning: Going beyond Euclidean data , 2016, IEEE Signal Process. Mag..

[38]  Yuval Tassa,et al.  Continuous control with deep reinforcement learning , 2015, ICLR.

[39]  Alán Aspuru-Guzik,et al.  Convolutional Networks on Graphs for Learning Molecular Fingerprints , 2015, NIPS.

[40]  Matt J. Kusner,et al.  Grammar Variational Autoencoder , 2017, ICML.

[41]  Niloy Ganguly,et al.  NeVAE: A Deep Generative Model for Molecular Graphs , 2018, AAAI.

[42]  Jean-Louis Reymond,et al.  Enumeration of 166 Billion Organic Small Molecules in the Chemical Universe Database GDB-17 , 2012, J. Chem. Inf. Model..

[43]  Max Welling,et al.  Semi-Supervised Classification with Graph Convolutional Networks , 2016, ICLR.

[44]  Thomas Demeester,et al.  Adversarial Sets for Regularising Neural Link Predictors , 2017, UAI.

[45]  Lantao Yu,et al.  SeqGAN: Sequence Generative Adversarial Nets with Policy Gradient , 2016, AAAI.

[46]  Alán Aspuru-Guzik,et al.  Objective-Reinforced Generative Adversarial Networks (ORGAN) for Sequence Generation Models , 2017, ArXiv.