Junction Tree Variational Autoencoder for Molecular Graph Generation

Abstract

We seek to automate the design of molecules based on specific chemical properties. In computational terms, this task involves continuous embedding and generation of molecular graphs. Our primary contribution is the direct realization of molecular graphs, a task previously approached by generating linear SMILES strings instead of graphs. Our junction tree variational autoencoder generates molecular graphs in two phases, by first generating a tree-structured scaffold over chemical substructures, and then combining them into a molecule with a graph message passing network. This approach allows us to incrementally expand molecules while maintaining chemical validity at every step. We evaluate our model on multiple tasks ranging from molecular generation to optimization. Across these tasks, our model outperforms previous state-of-the-art baselines by a significant margin.

References in corpus (9)

• Empirical Evaluation of Gated Recurrent Neural Networks on Sequence Modeling
• Semi-Supervised Classification with Graph Convolutional Networks
• Deep Convolutional Networks on Graph-Structured Data
• Learning Deep Generative Models of Graphs
• Syntax-Directed Variational Autoencoder for Structured Data
• Predicting Organic Reaction Outcomes with Weisfeiler-Lehman Network
• Grammar Variational Autoencoder
• Generating Focussed Molecule Libraries for Drug Discovery with Recurrent Neural Networks
• Actively Learning what makes a Discrete Sequence Valid

Cited by in corpus (171)

• On the Opportunities and Risks of Foundation Models
• Recent Advances and Applications of Deep Learning Methods in Materials Science
• Deep learning for molecular design - a review of the state of the art
• Deep Graph Library: A Graph-Centric, Highly-Performant Package for Graph Neural Networks
• Geometric Deep Learning: Grids, Groups, Graphs, Geodesics, and Gauges
• DeepAffinity: Interpretable Deep Learning of Compound-Protein Affinity through Unified Recurrent and Convolutional Neural Networks
• MolGAN: An implicit generative model for small molecular graphs
• Constrained Graph Variational Autoencoders for Molecule Design
• A Gentle Introduction to Deep Learning for Graphs
• Improving Graph Neural Network Expressivity via Subgraph Isomorphism Counting
• SELFIES and the future of molecular string representations
• Autonomous discovery in the chemical sciences part II: Outlook
• Graph Convolutional Policy Network for Goal-Directed Molecular Graph Generation
• Benchmarking Graph Neural Networks
• DGL-LifeSci: An Open-Source Toolkit for Deep Learning on Graphs in Life Science
• Conditional molecular design with deep generative models
• GraphAF: a Flow-based Autoregressive Model for Molecular Graph Generation
• Keeping it Simple: Language Models can learn Complex Molecular Distributions
• LanczosNet: Multi-Scale Deep Graph Convolutional Networks
• Molecular graph generation with Graph Neural Networks
• Design Space for Graph Neural Networks
• Graph-Bert: Only Attention is Needed for Learning Graph Representations
• Learning Multimodal Graph-to-Graph Translation for Molecular Optimization
• InfoGraph: Unsupervised and Semi-supervised Graph-Level Representation Learning via Mutual Information Maximization
• Graph Dynamical Networks for Unsupervised Learning of Atomic Scale Dynamics in Materials
• Machine learning modeling of family wide enzyme-substrate specificity screens
• Deep Learning and Knowledge-Based Methods for Computer Aided Molecular Design -- Toward a Unified Approach: State-of-the-Art and Future Directions
• Machine learning and AI-based approaches for bioactive ligand discovery and GPCR-ligand recognition
• Hierarchical Generation of Molecular Graphs using Structural Motifs
• Crystal Diffusion Variational Autoencoder for Periodic Material Generation
• From Variational to Deterministic Autoencoders
• A Deep Generative Model for Molecule Optimization via One Fragment Modification
• Generative Models for Automatic Chemical Design
• Multi-Objective Molecule Generation using Interpretable Substructures
• Machine Learning on Graphs: A Model and Comprehensive Taxonomy
• Augmenting Genetic Algorithms with Deep Neural Networks for Exploring the Chemical Space
• D-VAE: A Variational Autoencoder for Directed Acyclic Graphs
• Deeply learning molecular structure-property relationships using attention- and gate-augmented graph convolutional network
• Symmetry-adapted generation of 3d point sets for the targeted discovery of molecules
• Generative Chemical Transformer: Neural Machine Learning of Molecular Geometric Structures from Chemical Language via Attention
• Optimizing Molecules using Efficient Queries from Property Evaluations
• Learning Graph Models for Retrosynthesis Prediction
• Scaffold-constrained molecular generation
• Rethinking Graph Transformers with Spectral Attention
• Factorizable Graph Convolutional Networks
• DIG: A Turnkey Library for Diving into Graph Deep Learning Research
• Obey validity limits of data-driven models
• MolecularRNN: Generating realistic molecular graphs with optimized properties
• A Two-Step Graph Convolutional Decoder for Molecule Generation
• Rethinking pooling in graph neural networks
• We Should at Least Be Able to Design Molecules That Dock Well
• Learning to Make Generalizable and Diverse Predictions for Retrosynthesis
• Constrained Bayesian Optimization for Automatic Chemical Design
• Beyond Low-pass Filtering: Graph Convolutional Networks with Automatic Filtering
• Deep Molecular Dreaming: Inverse machine learning for de-novo molecular design and interpretability with surjective representations
• A Deep Generative Model for Fragment-Based Molecule Generation
• Bringing Your Own View: Graph Contrastive Learning without Prefabricated Data Augmentations
• Learning Gradient Fields for Molecular Conformation Generation
• Learning the Travelling Salesperson Problem Requires Rethinking Generalization
• All SMILES Variational Autoencoder
• Latent Molecular Optimization for Targeted Therapeutic Design
• Utilizing Reinforcement Learning for de novo Drug Design
• Sample-Efficient Optimization in the Latent Space of Deep Generative Models via Weighted Retraining
• Learning to Extend Molecular Scaffolds with Structural Motifs
• CogMol: Target-Specific and Selective Drug Design for COVID-19 Using Deep Generative Models
• Graphite: Iterative Generative Modeling of Graphs
• Simulated annealing for optimization of graphs and sequences
• Guiding Deep Molecular Optimization with Genetic Exploration
• A Tutorial on Deep Latent Variable Models of Natural Language
• Edge-based sequential graph generation with recurrent neural networks
• GraphDF: A Discrete Flow Model for Molecular Graph Generation
• Decoding Molecular Graph Embeddings with Reinforcement Learning
• Can Graph Neural Networks Count Substructures?
• Hierarchical Graph-to-Graph Translation for Molecules
• Active Learning for Graph Neural Networks via Node Feature Propagation
• Generating 3D Molecular Structures Conditional on a Receptor Binding Site with Deep Generative Models
• Subgroup Generalization and Fairness of Graph Neural Networks
• Graph Neural Networks with Learnable Structural and Positional Representations
• Data-Driven Approach to Encoding and Decoding 3-D Crystal Structures
• Learning Interpretable Representations of Entanglement in Quantum Optics Experiments using Deep Generative Models
• 3DMolNet: A Generative Network for Molecular Structures
• Improving black-box optimization in VAE latent space using decoder uncertainty
• Optimal Transport Graph Neural Networks
• Learning Sparse Interaction Graphs of Partially Detected Pedestrians for Trajectory Prediction
• A Model to Search for Synthesizable Molecules
• Differentiable Scaffolding Tree for Molecular Optimization
• JANUS: Parallel Tempered Genetic Algorithm Guided by Deep Neural Networks for Inverse Molecular Design
• Curiosity in exploring chemical space: Intrinsic rewards for deep molecular reinforcement learning
• Amortized Tree Generation for Bottom-up Synthesis Planning and Synthesizable Molecular Design
• Labeled Graph Generative Adversarial Networks
• PepCVAE: Semi-Supervised Targeted Design of Antimicrobial Peptide Sequences
• A generative model for molecule generation based on chemical reaction trees
• C5T5: Controllable Generation of Organic Molecules with Transformers
• Bayesian Variational Autoencoders for Unsupervised Out-of-Distribution Detection
• BatmanNet: Bi-branch Masked Graph Transformer Autoencoder for Molecular Representation
• Goal-directed graph construction using reinforcement learning
• Uncertainty quantification of molecular property prediction with Bayesian neural networks
• Categorical Normalizing Flows via Continuous Transformations
• Equivariant Subgraph Aggregation Networks
• Variational Autoencoder for Anti-Cancer Drug Response Prediction
• Score-based Generative Modeling in Latent Space
• Learning to simulate and design for structural engineering
• Towards Interpretable Sparse Graph Representation Learning with Laplacian Pooling
• MIMOSA: Multi-constraint Molecule Sampling for Molecule Optimization
• Advancing Biomedicine with Graph Representation Learning: Recent Progress, Challenges, and Future Directions
• Graph Transformation Policy Network for Chemical Reaction Prediction
• Learning a Continuous Representation of 3D Molecular Structures with Deep Generative Models
• Graph Generative Models for Fast Detector Simulations in High Energy Physics
• Adversarial Learned Molecular Graph Inference and Generation
• Directional Graph Networks
• A Variational-Sequential Graph Autoencoder for Neural Architecture Performance Prediction
• A Generative Model for Molecular Distance Geometry
• Practical Massively Parallel Monte-Carlo Tree Search Applied to Molecular Design
• Conditioned Generative Modeling of Molecular Glues: A Realistic AI Approach for Synthesizable Drug-like Molecules
• Unraveling the Potential of Diffusion Models in Small Molecule Generation
• Cascade-BGNN: Toward Efficient Self-supervised Representation Learning on Large-scale Bipartite Graphs
• Spatial Graph Attention and Curiosity-driven Policy for Antiviral Drug Discovery
• A 3D Generative Model for Structure-Based Drug Design
• Tiered Latent Representations and Latent Spaces for Molecular Graphs
• Graph-Aware Transformer: Is Attention All Graphs Need?
• Enhancing Model Learning and Interpretation Using Multiple Molecular Graph Representations for Compound Property and Activity Prediction
• HamNet: Conformation-Guided Molecular Representation with Hamiltonian Neural Networks
• Maximum Reward Formulation In Reinforcement Learning
• Graph Learning for Combinatorial Optimization: A Survey of State-of-the-Art
• Grammars and reinforcement learning for molecule optimization
• Chemical Structure Elucidation from Mass Spectrometry by Matching Substructures
• Graph convolutions that can finally model local structure
• DeepGG: a Deep Graph Generator
• Who Should Go First? A Self-Supervised Concept Sorting Model for Improving Taxonomy Expansion
• Probabilistic Generative Deep Learning for Molecular Design
• Building LEGO Using Deep Generative Models of Graphs
• Weisfeiler and Lehman Go Cellular: CW Networks
• Combining Latent Space and Structured Kernels for Bayesian Optimization over Combinatorial Spaces
• Direct Optimization through $\arg \max$ for Discrete Variational Auto-Encoder
• Attentional Multilabel Learning over Graphs: A Message Passing Approach
• Deep Denerative Models for Drug Design and Response
• Goal directed molecule generation using Monte Carlo Tree Search
• Controlled Molecule Generator for Optimizing Multiple Chemical Properties
• Physics-Constrained Predictive Molecular Latent Space Discovery with Graph Scattering Variational Autoencoder
• Graph Neural Distance Metric Learning with Graph-Bert
• Simulating Execution Time of Tensor Programs using Graph Neural Networks
• Discrete Object Generation with Reversible Inductive Construction
• Learning Space Partitions for Path Planning
• Permutation-Invariant Variational Autoencoder for Graph-Level Representation Learning
• Leveraging Active Subspaces to Capture Epistemic Model Uncertainty in Deep Generative Models for Molecular Design
• Uncertainty quantification of molecular property prediction using Bayesian neural network models
• Neural Trees for Learning on Graphs
• Enzyme promiscuity prediction using hierarchy-informed multi-label classification
• Benchmarking Deep Graph Generative Models for Optimizing New Drug Molecules for COVID-19
• Generative network complex (GNC) for drug discovery
• Partition and Code: learning how to compress graphs
• Vitruvion: A Generative Model of Parametric CAD Sketches
• Learning Markov Random Fields for Combinatorial Structures via Sampling through Lovász Local Lemma
• Molecular Attributes Transfer from Non-Parallel Data
• Contributions to Representation Learning with Graph Autoencoders and Applications to Music Recommendation
• TD-GEN: Graph Generation With Tree Decomposition
• Neighbor Embedding Variational Autoencoder
• Graph2Graph Learning with Conditional Autoregressive Models
• Analysis of training and seed bias in small molecules generated with a conditional graph-based variational autoencoder -- Insights for practical AI-driven molecule generation
• Self-learn to Explain Siamese Networks Robustly
• Probabilistic hypergraph grammars for efficient molecular optimization
• Adversarial Stein Training for Graph Energy Models
• Understanding in Artificial Intelligence
• High-Dimensional Bayesian Optimization via Semi-Supervised Learning with Optimized Unlabeled Data Sampling
• Generate Novel Molecules With Target Properties Using Conditional Generative Models
• Image-Like Graph Representations for Improved Molecular Property Prediction
• An Experimental Study of the Transferability of Spectral Graph Networks
• Evening the Score: Targeting SARS-CoV-2 Protease Inhibition in Graph Generative Models for Therapeutic Candidates
• Semantic Preserving Generative Adversarial Models
• Realistic molecule optimization on a learned graph manifold
• Visualizing Deep Graph Generative Models for Drug Discovery