mlsanityData to 14 Oct 2026

Improving Molecular Graph Neural Network Explainability with Orthonormalization and Induced Sparsity

arXiv 2105.04854 ↗First posted 11 May 2021stat.ML cs.LGcited by 20

Not on the map. Why?

Abstract (arXiv)

Rationalizing which parts of a molecule drive the predictions of a molecular graph convolutional neural network (GCNN) can be difficult. To help, we propose two simple regularization techniques to apply during the training of GCNNs: Batch Representation Orthonormalization (BRO) and Gini regularization. BRO, inspired by molecular orbital theory, encourages graph convolution operations to generate orthonormal node embeddings. Gini regularization is applied to the weights of the output layer and constrains the number of dimensions the model can use to make predictions. We show that Gini and BRO regularization can improve the accuracy of state-of-the-art GCNN attribution methods on artificial benchmark datasets. In a real-world setting, we demonstrate that medicinal chemists significantly prefer explanations extracted from regularized models. While we only study these regularizers in the context of GCNNs, both can be applied to other types of neural networks

Key facts. Cited by 20 papers in the map (data through 14 Oct 2026). Builds on 22 of its 47 references that are in the map. First posted on arXiv 11 May 2021.

Builds on

22 of 47 references.

References · 22

  1. 2014Adam: A Method for Stochastic Optimizationcited by 37,240
  2. 2016Grad-CAM: Visual Explanations from Deep Networks via Gradient-based Localizationcited by 6,733
  3. 2017Axiomatic Attribution for Deep Networkscited by 3,384
  4. 2015Learning Deep Features for Discriminative Localizationcited by 2,436
  5. 2019Fast Graph Representation Learning with PyTorch Geometriccited by 2,434
  6. 2017Learning Important Features Through Propagating Activation Differencescited by 1,458
  7. 2017SmoothGrad: removing noise by adding noisecited by 1,109
  8. 2018Deep Graph Infomaxcited by 1,106
  9. 2015Convolutional Networks on Graphs for Learning Molecular Fingerprintscited by 754
  10. 2019GNNExplainer: Generating Explanations for Graph Neural Networkscited by 737
  11. 2019InfoGraph: Unsupervised and Semi-supervised Graph-Level Representation Learning via Mutual Information Maximizationcited by 441
  12. 2015Unitary Evolution Recurrent Neural Networkscited by 294
10 more
  1. 2020Graph Information Bottleneckcited by 168
  2. 2017Interpretable Convolutional Neural Networkscited by 162
  3. 2019Explainability Techniques for Graph Convolutional Networkscited by 161
  4. 2017Interpreting Deep Visual Representations via Network Dissectioncited by 103
  5. 2017Orthogonal Weight Normalization: Solution to Optimization over Multiple Dependent Stiefel Manifolds in Deep Neural Networkscited by 92
  6. 2017On orthogonality and learning recurrent networks with long term dependenciescited by 83
  7. 2016Tunable Efficient Unitary Neural Networks (EUNN) and their application to RNNscited by 65
  8. 2017All You Need is Beyond a Good Init: Exploring Better Solution for Training Extremely Deep Convolutional Neural Networks with Orthonormality and Modulationcited by 58
  9. 2019Interpreting and Understanding Graph Convolutional Neural Network using Gradient-based Attribution Methodcited by 4
  10. 2020A comprehensive study on the prediction reliability of graph neural networks for virtual screeningcited by 1

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2022: 9; 2023: 4; 2024: 2; 2025: 3; 2026: 2.

20 papers.

Sort:

  1. 2022Interpretable and Generalizable Graph Learning via Stochastic Attention Mechanismcited by 171
  2. 2022Trustworthy Graph Neural Networks: Aspects, Methods and Trendscited by 89
  3. 2022Graph-based Molecular Representation Learningcited by 51
  4. 2023Encoding Time-Series Explanations through Self-Supervised Model Behavior Consistencycited by 42
  5. 2023A Survey on Explainability of Graph Neural Networkscited by 29
  6. 2022A Survey of Trustworthy Graph Learning: Reliability, Explainability, and Privacy Protectioncited by 20
  7. 2022Learning the Solution Operator of Boundary Value Problems using Graph Neural Networkscited by 17
  8. 2022MEGAN: Multi-Explanation Graph Attention Networkcited by 7
  9. 2022Explaining, Evaluating and Enhancing Neural Networks' Learned Representationscited by 6
  10. 2023Graph AI in Medicinecited by 3
  11. 2025Recent Developments in GNNs for Drug Discoverycited by 3
  12. 2022Physics-Informed Graph Learningcited by 2
8 more
  1. 2024Advancements in Molecular Property Prediction: A Survey of Single and Multimodal Approachescited by 2
  2. 2024Recent advances in interpretable machine learning using structure-based protein representationscited by 1
  3. 2025Fragment-Wise Interpretability in Graph Neural Networks via Molecule Decomposition and Contribution Analysiscited by 1
  4. 2022Towards Learning Self-Organized Criticality of Rydberg Atoms using Graph Neural Networkscited by 0
  5. 2023Streamlining models with explanations in the learning loopcited by 0
  6. 2025ARM-Explainer -- Explaining and improving graph neural network predictions for the maximum clique problem using node features and association rule miningcited by 0
  7. 2026Archetypal Graph Generative Models: Explainable and Identifiable Communities via Anchor-Dominant Convex Hullscited by 0
  8. 2026TAHB: A Comprehensive Benchmark for Text-Attributed Hypergraph Learningcited by 0