Constrained Bayesian Optimization for Automatic Chemical Design
arXiv:1709.05501
Abstract
Automatic Chemical Design is a framework for generating novel molecules with optimized properties. The original scheme, featuring Bayesian optimization over the latent space of a variational autoencoder, suffers from the pathology that it tends to produce invalid molecular structures. First, we demonstrate empirically that this pathology arises when the Bayesian optimization scheme queries latent points far away from the data on which the variational autoencoder has been trained. Secondly, by reformulating the search procedure as a constrained Bayesian optimization problem, we show that the effects of this pathology can be mitigated, yielding marked improvements in the validity of the generated molecules. We posit that constrained Bayesian optimization is a good approach for solving this class of training set mismatch in many generative tasks involving Bayesian optimization over the latent space of a variational autoencoder.
Previous versions accepted to the NIPS 2017 Workshop on Bayesian Optimization (BayesOpt 2017) and the NIPS 2017 Workshop on Machine Learning for Molecules and Materials
References in corpus (35)
- Practical Bayesian Optimization of Machine Learning Algorithms
- WaveNet: A Generative Model for Raw Audio
- Semi-Supervised Learning with Deep Generative Models
- Theano: new features and speed improvements
- DRAW: A Recurrent Neural Network For Image Generation
- Junction Tree Variational Autoencoder for Molecular Graph Generation
- Deep learning for molecular design - a review of the state of the art
- MolGAN: An implicit generative model for small molecular graphs
- Scalable Bayesian Optimization Using Deep Neural Networks
- Learning Deep Generative Models of Graphs
- Massively Multitask Networks for Drug Discovery
- Bayesian Optimization with Unknown Constraints
- ChemTS: An Efficient Python Library for de novo Molecular Generation
- Syntax-Directed Variational Autoencoder for Structured Data
- Conditional molecular design with deep generative models
- Improving Chemical Autoencoder Latent Space and Molecular De novo Generation Diversity with Heteroencoders
- Predicting Organic Reaction Outcomes with Weisfeiler-Lehman Network
- Bayesian semi-supervised learning for uncertainty-calibrated prediction of molecular properties and active learning
- Deep Confidence: A Computationally Efficient Framework for Calculating Reliable Errors for Deep Neural Networks
- Generative Models for Automatic Chemical Design
- Deeply learning molecular structure-property relationships using attention- and gate-augmented graph convolutional network
- ChemBO: Bayesian Optimization of Small Organic Molecules with Synthesizable Recommendations
- MolecularRNN: Generating realistic molecular graphs with optimized properties
- Molecular Hypergraph Grammar with its Application to Molecular Optimization
- An Entropy Search Portfolio for Bayesian Optimization
- Latent Molecular Optimization for Targeted Therapeutic Design
- In silico generation of novel, drug-like chemical matter using the LSTM neural network
- Graph Convolutional Neural Networks for Polymers Property Prediction
- A Model to Search for Synthesizable Molecules
- Actively Learning what makes a Discrete Sequence Valid
- Uncertainty quantification of molecular property prediction with Bayesian neural networks
- A COLD Approach to Generating Optimal Samples
- Probabilistic Generative Deep Learning for Molecular Design
- Grammars and reinforcement learning for molecule optimization
- Multiple-objective Reinforcement Learning for Inverse Design and Identification
Cited by in corpus (24)
- Deep learning for molecular design - a review of the state of the art
- ChemBO: Bayesian Optimization of Small Organic Molecules with Synthesizable Recommendations
- Tuning Hyperparameters without Grad Students: Scalable and Robust Bayesian Optimisation with Dragonfly
- Latent Molecular Optimization for Targeted Therapeutic Design
- Generative Multi-Form Bayesian Optimization
- BOSS: Bayesian Optimization over String Spaces
- Noisy-Input Entropy Search for Efficient Robust Bayesian Optimization
- Bayesian Variational Autoencoders for Unsupervised Out-of-Distribution Detection
- GIBBON: General-purpose Information-Based Bayesian OptimisatioN
- Achieving Robustness to Aleatoric Uncertainty with Heteroscedastic Bayesian Optimisation
- AutoOED: Automated Optimal Experiment Design Platform
- Kernels over Sets of Finite Sets using RKHS Embeddings, with Application to Bayesian (Combinatorial) Optimization
- A COLD Approach to Generating Optimal Samples
- Computationally Efficient High-Dimensional Bayesian Optimization via Variable Selection
- Leveraging binding-site structure for drug discovery with point-cloud methods
- Scalable Thompson Sampling using Sparse Gaussian Process Models
- High-dimensional Black-box Optimization Under Uncertainty
- Efficient Transfer Bayesian Optimization with Auxiliary Information
- Emergent Hand Morphology and Control from Optimizing Robust Grasps of Diverse Objects
- Good practices for Bayesian Optimization of high dimensional structured spaces
- Generative network complex (GNC) for drug discovery
- End-to-End Learning of Deep Kernel Acquisition Functions for Bayesian Optimization
- Classified Regression for Bayesian Optimization: Robot Learning with Unknown Penalties
- Sonic: A Sampling-based Online Controller for Streaming Applications