Sample-Efficient Optimization in the Latent Space of Deep Generative Models via Weighted Retraining

Abstract

Many important problems in science and engineering, such as drug design, involve optimizing an expensive black-box objective function over a complex, high-dimensional, and structured input space. Although machine learning techniques have shown promise in solving such problems, existing approaches substantially lack sample efficiency. We introduce an improved method for efficient black-box optimization, which performs the optimization in the low-dimensional, continuous latent manifold learned by a deep generative model. In contrast to previous approaches, we actively steer the generative model to maintain a latent manifold that is highly useful for efficiently optimizing the objective. We achieve this by periodically retraining the generative model on the data points queried along the optimization trajectory, as well as weighting those data points according to their objective function value. This weighted retraining can be easily implemented on top of existing methods, and is empirically shown to significantly improve their efficiency and performance on synthetic and real-world optimization problems.

arXiv comment23 pages, 14 figures; Includes supplementary material; NeurIPS 2020

References in corpus (23)

• Conditional Generative Adversarial Nets
• Practical Bayesian Optimization of Machine Learning Algorithms
• A Tutorial on Bayesian Optimization of Expensive Cost Functions, with Application to Active User Modeling and Hierarchical Reinforcement Learning
• Conditional Image Generation with PixelCNN Decoders
• Neural Architecture Search: A Survey
• Junction Tree Variational Autoencoder for Molecular Graph Generation
• Deep learning for molecular design - a review of the state of the art
• Objective-Reinforced Generative Adversarial Networks (ORGAN) for Sequence Generation Models
• Neural Architecture Optimization
• MolGAN: An implicit generative model for small molecular graphs
• Syntax-Directed Variational Autoencoder for Structured Data
• Conditional molecular design with deep generative models
• Learning Multimodal Graph-to-Graph Translation for Molecular Optimization
• Bayesian Optimization of Combinatorial Structures
• ChemBO: Bayesian Optimization of Small Organic Molecules with Synthesizable Recommendations
• Molecular Hypergraph Grammar with its Application to Molecular Optimization
• Getting a CLUE: A Method for Explaining Uncertainty Estimates
• Design by adaptive sampling
• Active Learning of Linear Embeddings for Gaussian Processes
• Decentralized High-Dimensional Bayesian Optimization with Factor Graphs
• Bayesian Variational Autoencoders for Unsupervised Out-of-Distribution Detection
• A COLD Approach to Generating Optimal Samples
• Bayesian Optimization with Approximate Set Kernels

Cited by in corpus (8)

• Gryffin: An algorithm for Bayesian optimization of categorical variables informed by expert knowledge
• Multi-Objective Latent Space Optimization of Generative Molecular Design Models
• Gaussian Process Molecule Property Prediction with FlowMO
• Generative Multi-Form Bayesian Optimization
• Bayesian Variational Autoencoders for Unsupervised Out-of-Distribution Detection
• CircuitVAE: Efficient and Scalable Latent Circuit Optimization
• Leveraging Active Subspaces to Capture Epistemic Model Uncertainty in Deep Generative Models for Molecular Design
• High-Dimensional Bayesian Optimization via Semi-Supervised Learning with Optimized Unlabeled Data Sampling