Improving Molecular Design by Stochastic Iterative Target Augmentation
This work addresses the problem of limited training data for researchers and practitioners in molecular design, offering an incremental improvement through a novel self-training method.
The paper tackles the challenge of training data scarcity in generative models for molecular design by introducing a self-training approach that iteratively creates additional molecular targets, resulting in a 10% absolute gain over previous state-of-the-art methods in conditional molecular design.
Generative models in molecular design tend to be richly parameterized, data-hungry neural models, as they must create complex structured objects as outputs. Estimating such models from data may be challenging due to the lack of sufficient training data. In this paper, we propose a surprisingly effective self-training approach for iteratively creating additional molecular targets. We first pre-train the generative model together with a simple property predictor. The property predictor is then used as a likelihood model for filtering candidate structures from the generative model. Additional targets are iteratively produced and used in the course of stochastic EM iterations to maximize the log-likelihood that the candidate structures are accepted. A simple rejection (re-weighting) sampler suffices to draw posterior samples since the generative model is already reasonable after pre-training. We demonstrate significant gains over strong baselines for both unconditional and conditional molecular design. In particular, our approach outperforms the previous state-of-the-art in conditional molecular design by over 10% in absolute gain. Finally, we show that our approach is useful in other domains as well, such as program synthesis.