Improving black-box optimization in VAE latent space using decoder uncertaintyDownload PDF

21 May 2021, 20:43 (edited 08 Nov 2021)NeurIPS 2021 PosterReaders: Everyone
  • Keywords: Optimization, Uncertainty, High dimensionality, Discrete data, Variational autoencoder, Generative models, Molecule generation
  • TL;DR: We improve the black-box optimization of high-dimensional discrete objects with variational autoencoders by leveraging the epistemic uncertainty of the decoder to avoid regions of latent space leading to low-quality decodings.
  • Abstract: Optimization in the latent space of variational autoencoders is a promising approach to generate high-dimensional discrete objects that maximize an expensive black-box property (e.g., drug-likeness in molecular generation, function approximation with arithmetic expressions). However, existing methods lack robustness as they may decide to explore areas of the latent space for which no data was available during training and where the decoder can be unreliable, leading to the generation of unrealistic or invalid objects. We propose to leverage the epistemic uncertainty of the decoder to guide the optimization process. This is not trivial though, as a naive estimation of uncertainty in the high-dimensional and structured settings we consider would result in high estimator variance. To solve this problem, we introduce an importance sampling-based estimator that provides more robust estimates of epistemic uncertainty. Our uncertainty-guided optimization approach does not require modifications of the model architecture nor the training process. It produces samples with a better trade-off between black-box objective and validity of the generated samples, sometimes improving both simultaneously. We illustrate these advantages across several experimental settings in digit generation, arithmetic expression approximation and molecule generation for drug design.
  • Supplementary Material: pdf
  • Code Of Conduct: I certify that all co-authors of this work have read and commit to adhering to the NeurIPS Statement on Ethics, Fairness, Inclusivity, and Code of Conduct.
  • Code:
17 Replies