Constrained Generation of Semantically Valid Graphs via Regularizing Variational Autoencoders

Abstract

Deep generative models have achieved remarkable success in various data domains, including images, time series, and natural languages. There remain, however, substantial challenges for combinatorial structures, including graphs. One of the key challenges lies in the difficulty of ensuring semantic validity in context. For examples, in molecular graphs, the number of bonding-electron pairs must not exceed the valence of an atom; whereas in protein interaction networks, two proteins may be connected only when they belong to the same or correlated gene ontology terms. These constraints are not easy to be incorporated into a generative model. In this work, we propose a regularization framework for variational autoencoders as a step toward semantic validity. We focus on the matrix representation of graphs and formulate penalty terms that regularize the output distribution of the decoder to encourage the satisfaction of validity constraints. Experimental results confirm a much higher likelihood of sampling valid graphs in our approach, compared with others reported in the literature.

NIPS 2018

Cited by in corpus (30)

• Graph Neural Networks: A Review of Methods and Applications
• Deep learning for molecular design - a review of the state of the art
• Self-Referencing Embedded Strings (SELFIES): A 100% robust molecular string representation
• GraphAF: a Flow-based Autoregressive Model for Molecular Graph Generation
• Hierarchical Generation of Molecular Graphs using Structural Motifs
• Multi-Objective Molecule Generation using Interpretable Substructures
• D-VAE: A Variational Autoencoder for Directed Acyclic Graphs
• A Deep Generative Model for Graph Layout
• Deep Molecular Dreaming: Inverse machine learning for de-novo molecular design and interpretability with surjective representations
• All SMILES Variational Autoencoder
• CogMol: Target-Specific and Selective Drug Design for COVID-19 Using Deep Generative Models
• Graph Residual Flow for Molecular Graph Generation
• Hierarchical Graph-to-Graph Translation for Molecules
• Controllable Data Generation by Deep Learning: A Review
• Keep It Simple: Graph Autoencoders Without Graph Convolutional Networks
• Categorical Normalizing Flows via Continuous Transformations
• Graph Coarsening with Neural Networks
• Adversarial Learned Molecular Graph Inference and Generation
• Probabilistic Generative Deep Learning for Molecular Design
• A Systematic Assessment of Deep Learning Models for Molecule Generation
• Auto-decoding Graphs
• Generating a Doppelganger Graph: Resembling but Distinct
• SHADOWCAST: Controllable Graph Generation
• Graph Energy-based Model for Substructure Preserving Molecular Design
• Factorized Deep Generative Models for Trajectory Generation with Spatiotemporal-Validity Constraints
• Operator Autoencoders: Learning Physical Operations on Encoded Molecular Graphs
• Learning to Sample Hard Instances for Graph Algorithms
• That looks interesting! Personalizing Communication and Segmentation with Random Forest Node Embeddings
• Analysis of training and seed bias in small molecules generated with a conditional graph-based variational autoencoder -- Insights for practical AI-driven molecule generation
• Generating the Graph Gestalt: Kernel-Regularized Graph Representation Learning