DropConn: Dropout Connection Based Random GNNs for Molecular Property Prediction

Recently, molecular data mining has attracted a lot of attention owing to its great application potential in material and drug discovery. However, this mining task faces a challenge posed by the scarcity of labeled molecular graphs. To overcome this challenge, we introduce a novel data augmentation and a semi-supervised confidence-aware consistency regularization training framework for molecular property prediction. The core of our framework is a data augmentation strategy on molecular graphs, named DropConn (Dropout Connection). DropConn generates pseudo molecular graphs by softening the hard connections of chemical bonds (as edges), where the soft weights are calculated from edge features so that the adaptive interactions between different atoms can be incorporated. Besides, to enhance the model's generalization ability, a consistency regularization training strategy is proposed to take full advantage of massive unlabeled data. Furthermore, DropConn can serve as a plugin that can be seamlessly added to many existing models. Extensive experiments under both non-pre-training setting and fine-tuning setting demonstrate that DropConn can obtain superior performance (up to 8.22%) over state-of-the-art methods on molecular property prediction tasks.

• Publication year

  2024

• Venue

  IEEE Transactions on Knowledge and Data Engineering

• Publication date

  2024-02-01

• Fields of study

  Chemistry, Materials Science, Computer Science

• Identifiers
  DOI 10.1109/TKDE.2023.3290032
• External record

  Open on Semantic Scholar

• Source metadata

  Semantic Scholar

• No claims are published for this paper.

• No concepts are published for this paper.

• GRAND+: Scalable Graph Random Neural Networks

  2022cited by this paper
• Interpretable and Generalizable Graph Learning via Stochastic Attention Mechanism

  2022cited by this paper
• KPGT: Knowledge-Guided Pre-training of Graph Transformer for Molecular Property Prediction

  2022cited by this paper
• GraphMAE: Self-Supervised Masked Graph Autoencoders

  2022cited by this paper
• Mixup for Node and Graph Classification

  2021cited by this paper
• Rethinking Graph Transformers with Spectral Attention

  2021cited by this paper
• Do Transformers Really Perform Badly for Graph Representation?

  2021cited by this paper
• CogDL: A Comprehensive Library for Graph Deep Learning

  2021cited by this paper
• Graph Structure Learning with Variational Information Bottleneck

  2021cited by this paper
• Machine learning for chemical discovery

  2020cited by this paper
• Principal Neighbourhood Aggregation for Graph Nets

  2020cited by this paper
• DeeperGCN: All You Need to Train Deeper GCNs

  2020cited by this paper
• GraphNorm: A Principled Approach to Accelerating Graph Neural Network Training

  2020cited by this paper
• Directional Graph Networks

  2020cited by this paper
• GraphMix: Improved Training of GNNs for Semi-Supervised Learning

  2020cited by this paper
• Graph Random Neural Networks for Semi-Supervised Learning on Graphs

  2020influential reference
• Graph convolutions that can finally model local structure

  2020cited by this paper
• Graph Stochastic Neural Networks for Semi-supervised Learning

  2020cited by this paper
• A Generalization of Transformer Networks to Graphs

  2020cited by this paper
• Robust Optimization as Data Augmentation for Large-scale Graphs

  2020cited by this paper
• Strategies for Pre-training Graph Neural Networks

  2019cited by this paper
• MixMatch: A Holistic Approach to Semi-Supervised Learning

  2019cited by this paper
• InfoGraph: Unsupervised and Semi-supervised Graph-Level Representation Learning via Mutual Information Maximization

  2019cited by this paper
• Measuring and Relieving the Over-smoothing Problem for Graph Neural Networks from the Topological View

  2019cited by this paper
• DropEdge: Towards Deep Graph Convolutional Networks on Node Classification

  2019influential reference
• Predict then Propagate: Graph Neural Networks meet Personalized PageRank

  2018cited by this paper
• Hierarchical Graph Representation Learning with Differentiable Pooling

  2018cited by this paper
• Attention-based Graph Neural Network for Semi-supervised Learning

  2018cited by this paper
• Representation Learning on Graphs with Jumping Knowledge Networks

  2018cited by this paper
• A simple neural network module for relational reasoning

  2017cited by this paper
• Inductive Representation Learning on Large Graphs

  2017cited by this paper
• Predicting multicellular function through multi-layer tissue networks

  2017cited by this paper
• Gated Graph Sequence Neural Networks

  2015cited by this paper
• Adam: A Method for Stochastic Optimization

  2014influential reference
• Comparison of descriptor spaces for chemical compound retrieval and classification

  2006cited by this paper

• Network Measure-Enriched GNNs: A New Framework for Power Grid Stability Prediction

  2026cites this paper
• Endowing Molecular Language with Geometry Perception via Modality Compensation for High-Throughput Quantum Hamiltonian Prediction

  2026cites this paper
• Consistency-regularized graph neural networks for molecular property prediction

  2025cites this paper
• Federated Prompt Learning for Few-Shot Graph Classification

  2025cites this paper
• Higher-order Structure and Semantics-enhanced User Profiling for Recommendation

  2025cites this paper
• A Data-Driven Scale-Adaptive Time-Frequency Convolutional Network for Long Sequence Time-Series Forecasting

  2025cites this paper
• Adaptively multi-modal contrastive fusion network for molecular properties prediction

  2025cites this paper
• Fragment-Masked Diffusion for Molecular Optimization

  2024cites this paper
• Impact of Domain Knowledge and Multi-Modality on Intelligent Molecular Property Prediction: A Systematic Survey

  2024cites this paper
• Mix-Key: graph mixup with key structures for molecular property prediction

  2024cites this paper
• Graph Rewiring and Preprocessing for Graph Neural Networks Based on Effective Resistance

  2024cites this paper
• LGB: Language Model and Graph Neural Network-Driven Social Bot Detection

  2024cites this paper
• Effective Entry-Wise Flow for Molecule Generation

  2024cites this paper
• Generalized Graph Prompt: Toward a Unification of Pre-Training and Downstream Tasks on Graphs

  2023cites this paper
• Bridging the Semantic-Numerical Gap: A Numerical Reasoning Method of Cross-modal Knowledge Graph for Material Property Prediction

  2023cites this paper