Author(s)

Lukas Gruber

Affiliation(s)

University of Vienna, Vienna 1010, Austria.

Corresponding Author

Lukas Gruber

ABSTRACT

Molecular graph representation learning has emerged as a fundamental challenge in computational chemistry and drug discovery, where the ability to capture multi-scale structural information from atomic to molecular levels is crucial for accurate property prediction and molecular design. Traditional graph neural network approaches typically operate at a single resolution, failing to capture the hierarchical nature of molecular structure that spans multiple geometric scales from local atomic environments to global molecular topology. This paper presents a novel framework that employs Adaptive Molecular Graph Abstraction (AMGA) through reinforcement learning to dynamically learn optimal abstraction strategies across different geometric scales. The proposed framework combines hierarchical graph pooling mechanisms with reinforcement learning agents that adaptively select abstraction levels and pooling strategies based on molecular characteristics and task requirements. Our approach employs differentiable clustering algorithms that learn to group atoms into chemically meaningful motifs, progressing through multiple pooling levels to create increasingly abstract molecular representations. The reinforcement learning component formulates abstraction strategy selection as a sequential decision-making problem, enabling dynamic adaptation to different molecular families and property types. Experimental evaluation demonstrates superior reconstruction accuracy compared to existing approaches, with our method maintaining over 85% accuracy across all molecule sizes while baseline methods like CG-VAE deteriorate significantly for larger molecules. The adaptive abstraction mechanism enables automatic discovery of optimal representation granularities for different chemical contexts, providing interpretable insights into structure-property relationships while maintaining computational efficiency suitable for large-scale molecular screening applications.

KEYWORDS

Molecular graph abstraction; Reinforcement learning; Hierarchical pooling; Graph neural networks; Molecular property prediction; Multi-scale representation

CITE THIS PAPER

Lukas Gruber. Adaptive molecular graph abstraction via reinforcement learning across geometric scales. World Journal of Materials Science.  2025, 3(1): 20-28. DOI: https://doi.org/10.61784/wjms3010.

REFERENCES

[1] Wigh DS, Goodman JM, Lapkin AA. A review of molecular representation in the age of machine learning. Wiley Interdisciplinary Reviews: Computational Molecular Science. 2022, 12(5): e1603.

[2] Elton DC, Boukouvalas Z, Fuge MD, et al. Deep learning for molecular design-a review of the state of the art. Molecular Systems Design & Engineering. 2019, 4(4): 828-849.

[3] Shen J, Nicolaou CA. Molecular property prediction: recent trends in the era of artificial intelligence. Drug Discovery Today: Technologies. 2019, 32: 29-36.

[4] Pereira TO. Deep learning in drug development: Optimizing de novo molecular design with deep reinforcement learning. Universidade de Coimbra. 2025. (Doctoral dissertation).

[5] Wellhausen L, Hutter M. Rough terrain navigation for legged robots using reachability planning and template learning. 2021 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). 2021: 6914-6921.

[6] Hu X, Zhao X, Wang J, et al. Information-theoretic multi-scale geometric pre-training for enhanced molecular property prediction. PLOS One. 2025.

[7] Sheshanarayana R, You F. Molecular representation learning: cross-domain foundations and future frontiers. Digital Discovery. 2025.

[8] Reiser P, Neubert M, Eberhard A, et al. Graph neural networks for materials science and chemistry. Communications Materials. 2022, 3(1): 93.

[9] Liyaqat T, Ahmad T, Saxena C. Advancements in molecular property prediction: a survey of single and multimodal approaches. Archives of Computational Methods in Engineering. 2025: 1-31.

[10] Trifan A, Gorgun D, Salim M, et al. Intelligent resolution: Integrating Cryo-EM with AI-driven multi-resolution simulations to observe the severe acute respiratory syndrome coronavirus-2 replication-transcription machinery in action. The International Journal of High Performance Computing Applications. 2022, 36(5-6): 603-623.

[11] Leite LS, Banerjee S, Wei Y, et al. Modern chemical graph theory. Wiley Interdisciplinary Reviews: Computational Molecular Science. 2024, 14(5): e1729.

[12] Mondonico L, Su H, Florian J, et al. Hierarchical moiety-aware graph transformer for Li-metal electrolyte formulation design. 2025.

[13] Walters WP, Barzilay R. Applications of deep learning in molecule generation and molecular property prediction. Accounts of Chemical Research. 2020, 54(2): 263-270.

[14] Munikoti S, Agarwal D, Das L, et al. Challenges and opportunities in deep reinforcement learning with graph neural networks: A comprehensive review of algorithms and applications. IEEE Transactions on Neural Networks and Learning Systems. 2023, 35(11): 15051-15071.

[15] Iranfar A, Zapater M, Atienza D. Multiagent reinforcement learning for hyperparameter optimization of convolutional neural networks. IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems. 2021, 41(4): 1034-1047.

[16] David L, Thakkar A, Mercado R, et al. Molecular representations in AI-driven drug discovery: a review and practical guide. Journal of Cheminformatics. 2020, 12(1): 56.

[17] Chen S, Liu Y, Zhang Q, et al. Multi-distance spatial-temporal graph neural network for anomaly detection in blockchain transactions. Advanced Intelligent Systems. 2025: 2400898.

[18] Zhang X, Chen S, Shao Z, et al. Enhanced lithographic hotspot detection via multi-task deep learning with synthetic pattern generation. IEEE Open Journal of the Computer Society. 2024.

[19] Zhang Q, Chen S, Liu W. Balanced knowledge transfer in MTTL-ClinicalBERT: A symmetrical multi-task learning framework for clinical text classification. Symmetry. 2025, 17(6): 823.

[20] Shao Z, Wang X, Ji E, et al. GNN-EADD: Graph neural network-based e-commerce anomaly detection via dual-stage learning. IEEE Access. 2025.

[21] Li P, Ren S, Zhang Q, et al. Think4SCND: Reinforcement learning with thinking model for dynamic supply chain network design. IEEE Access. 2024.

[22] Ren S, Jin J, Niu G, et al. ARCS: Adaptive reinforcement learning framework for automated cybersecurity incident response strategy optimization. Applied Sciences. 2025, 15(2): 951.

[23] Cao J, Zheng W, Ge Y, et al. DriftShield: Autonomous fraud detection via actor-critic reinforcement learning with dynamic feature reweighting. IEEE Open Journal of the Computer Society. 2025.

[24] Wang J, Liu J, Zheng W, et al. Temporal heterogeneous graph contrastive learning for fraud detection in credit card transactions. IEEE Access. 2025.

[25] Mai NT, Cao W, Liu W. Interpretable knowledge tracing via transformer-Bayesian hybrid networks: Learning temporal dependencies and causal structures in educational data. Applied Sciences. 2025, 15(17): 9605.

[26] Cao W, Mai NT, Liu W. Adaptive knowledge assessment via symmetric hierarchical Bayesian neural networks with graph symmetry-aware concept dependencies. Symmetry. 2025, 17(8): 1332.

[27] Mai NT, Cao W, Wang Y. The global belonging support framework: Enhancing equity and access for international graduate students. Journal of International Students. 2025, 15(9): 141-160.

[28] Tan Y, Wu B, Cao J, et al. LLaMA-UTP: Knowledge-guided expert mixture for analyzing uncertain tax positions. IEEE Access. 2025.

[29] Sun T, Yang J, Li J, et al. Enhancing auto insurance risk evaluation with transformer and SHAP. IEEE Access. 2024.

[30] Zhang H, Ge Y, Zhao X, et al. Hierarchical deep reinforcement learning for multi-objective integrated circuit physical layout optimization with congestion-aware reward shaping. IEEE Access. 2025.

[31] Ji E, Wang Y, Xing S, et al. Hierarchical reinforcement learning for energy-efficient API traffic optimization in large-scale advertising systems. IEEE Access. 2025.

[32] Jin J, Xing S, Ji E, et al. XGate: Explainable reinforcement learning for transparent and trustworthy API traffic management in IoT sensor networks. Sensors. 2025, 25(7): 2183.