mlsanityData to 14 Oct 2026

Global Human-guided Counterfactual Explanations for Molecular Properties via Reinforcement Learning

arXiv 2406.13869 ↗First posted 19 Jun 2024cs.LG q-bio.BMcited by 1

On the map: Deep reinforcement learning methods → RL for energy, networks and operations

Abstract (arXiv)

Counterfactual explanations of Graph Neural Networks (GNNs) offer a powerful way to understand data that can naturally be represented by a graph structure. Furthermore, in many domains, it is highly desirable to derive data-driven global explanations or rules that can better explain the high-level properties of the models and data in question. However, evaluating global counterfactual explanations is hard in real-world datasets due to a lack of human-annotated ground truth, which limits their use in areas like molecular sciences. Additionally, the increasing scale of these datasets provides a challenge for random search-based methods. In this paper, we develop a novel global explanation model RLHEX for molecular property prediction. It aligns the counterfactual explanations with human-defined principles, making the explanations more interpretable and easy for experts to evaluate. RLHEX includes a VAE-based graph generator to generate global explanations and an adapter to adjust the latent representation space to human-defined principles. Optimized by Proximal Policy Optimization (PPO), the global explanations produced by RLHEX cover 4.12% more input graphs and reduce the distance between the counterfactual explanation set and the input set by 0.47% on average across three molecular datasets. RLHEX provides a flexible framework to incorporate different human-designed principles into the counterfactual explanation generation process, aligning these explanations with domain expertise. The code and data are released at https://github.com/dqwang122/RLHEX.

Key facts. Cited by 1 paper in the map (data through 14 Oct 2026). Placed in RL for energy, networks and operations, one of 15 directions in Deep reinforcement learning methods. Builds on 27 of its 52 references that are in the map. First posted on arXiv 19 Jun 2024.

Builds on

27 of 52 references.

References · 27

  1. 2017Attention Is All You Needcited by 60,361
  2. 2014Adam: A Method for Stochastic Optimizationcited by 37,240
  3. 2017Proximal Policy Optimization Algorithmscited by 15,215
  4. 2017Deep reinforcement learning from human preferencescited by 3,912
  5. 2022Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedbackcited by 3,720
  6. 2017Neural Message Passing for Quantum Chemistrycited by 3,487
  7. 2022Constitutional AI: Harmlessness from AI Feedbackcited by 3,050
  8. 2019Fine-Tuning Language Models from Human Preferencescited by 2,150
  9. 2018Graph Neural Networks: A Review of Methods and Applicationscited by 1,576
  10. 2016Automatic chemical design using a data-driven continuous representation of moleculescited by 765
  11. 2018Junction Tree Variational Autoencoder for Molecular Graph Generationcited by 652
  12. 2018GraphVAE: Towards Generation of Small Graphs Using Variational Autoencoderscited by 383
15 more
  1. 2018Graph Convolutional Policy Network for Goal-Directed Molecular Graph Generationcited by 364
  2. 2023Principle-Driven Self-Alignment of Language Models from Scratch with Minimal Human Supervisioncited by 318
  3. 2018Learning Deep Generative Models of Graphscited by 229
  4. 2020XGNN: Towards Model-Level Explanations of Graph Neural Networkscited by 211
  5. 2021MARS: Markov Molecular Sampling for Multi-objective Drug Discoverycited by 125
  6. 2022Learning and Evaluating Graph Neural Network Explanations based on Counterfactual and Factual Reasoningcited by 79
  7. 2021Robust Counterfactual Explanations on Graph Neural Networkscited by 72
  8. 2021Molecule Generation by Principal Subgraph Mining and Assemblingcited by 54
  9. 2022CLEAR: Generative Counterfactual Explanations on Graphscited by 48
  10. 2023De Novo Molecular Generation via Connection-aware Motif Miningcited by 46
  11. 2021Counterfactual Graphs for Explainable Classification of Brain Networkscited by 33
  12. 2021MEG: Generating Molecular Counterfactual Explanations for Deep Graph Networkscited by 32
  13. 2023Selective Explanations: Leveraging Human Input to Align Explainable AIcited by 29
  14. 2022MolE: a molecular foundation model for drug discoverycited by 12
  15. 2023GNNX-BENCH: Unravelling the Utility of Perturbation-based GNN Explainers through In-depth Benchmarkingcited by 11

Built on by

Cited by year

Citing papers per year: 2024: 1.

1 paper.

  1. 2024Ranking over Regression for Bayesian Optimization and Molecule Selectioncited by 4

Similar, not cited

Close in wording, no citation either way

  1. 2026Faithful, Sufficient and Understandable: Rethinking Graph Counterfactual Explanations via Discrete Diffusion Inversion
  2. 2024Explainable Molecular Property Prediction: Aligning Chemical Concepts with Predictions via Language Models
  3. 2025Reasoning-Enhanced Large Language Models for Molecular Property Prediction
  4. 2026Improving Molecular Property Prediction in Small Language Models Using Graph-based Tools
  5. 2025Adaptive Substructure-Aware Expert Model for Molecular Property Prediction
  6. 2025MolGraph-xLSTM: A graph-based dual-level xLSTM framework with multi-head mixture-of-experts for enhanced molecular representation and interpretability
  7. 2024LLM and GNN are Complementary: Distilling LLM for Multimodal Graph Learning
  8. 2024MolecularGPT: Open Large Language Model (LLM) for Few-Shot Molecular Property Prediction