Concat-tree genetic programming for graph neural architecture search: Efficient materials property classification toward high-throughput screening
Maciej Krzywda*1
, Mariusz Wermiński2, Szymon Łukasik1, Amir H. Gandomi3
1AGH University of Krakow, Faculty of Physics and Applied Computer Science, al. A. Mickiewicza 30, 30-059 Krakow, Poland.
2AGH University of Krakow, Faculty of Metals Engineering and Industrial Computer Science, al. A. Mickiewicza 30, 30-059 Krakow, Poland.
3Faculty of Engineering and IT, University of Technology Sydney, 5 Broadway, Ultimo NSW 2007, Australia.
DOI:
https://doi.org/10.7494/cmms.2026.4.1050
Abstract:
We propose Concat-Tree Genetic Programming (CTGP) to automatically discover graph neural network (GNN) architectures for binary classification of DFT-derived materials properties in the JARVIS-DFT database. CTGP is inspired by Cartesian genetic programming (CGP) and classical tree-based genetic programming (GP): genotypes are binary concat trees that decode into explicit, heterogeneous sequential GNN stacks, evolved by mutation-only elitist search. Seven tasks span electronic structure (metal vs. insulator), magnetism, photovoltaic efficiency proxies, topological band character, and thermoelectric transport. Using fixed JARVIS train/validation/test splits and published binarization thresholds, we compare four search configurations – (1 + 1) and (1 + 4) regimes, each with and without a learned surrogate filter – with MLP, RF, DT, and SVM baselines and fair GNN references across four representation levels (L0–L3).L0 uses atomic number only, L1 adds learned elemental descriptors, L2 adds bond-distance summaries, and L3 adds bond-angle statistics. Each configuration uses ten independent random seeds per task. The search yields compact architectures with tens to hundreds of thousands of parameters and millisecond-scale model-forward inference. Accuracy is reported with model complexity, search cost, test AUC, macro-F1, and property-aware analysis of discovered operator sequences. Fixed GNN baselines and random search often match or exceed CTGP on test accuracy, while CTGP remains competitive on selected electronic-structure tasks and provides architecturally transparent stacks for exploratory, property-related interpretation, without claiming causal explainability. On benchmarks with a stable signal above majority-class performance, these surrogates may support rapid screening of large preprocessed graph libraries at low inference cost.
Cite as:
Krzywda1*, M., Wermiński, M., Łukasik, S., & Gandomi, A.H. (2026). Concat-tree genetic programming for graph neural architecture search: Efficient materials property classification toward high-throughput screening. Computer Methods in Materials Science, 26(4), XX-XX. https://doi.org/10.7494/cmms.2026.3.1050
Article (PDF):
Accepted manuscript: the final PDF will be available soon.
Keywords:
graph neural networks, genetic programming, Cartesian genetic programming, concat-tree genetic programming, structure-property relationships surrogate methods, structure-property relationships
Publication dates:
Received: 31.07.2026, Accepted: 08.09.2026, Published: XX.11.2026
Publication type:
Original scientific paper
References:
Bhattacharya, B., Dinesh Kumar, G. R., Agarwal, A., Erkoç, Ş., Singh, A., & Chakraborti, N. (2009). Analyzing Fe–Zn system using molecular dynamics, evolutionary neural nets and multi-objective genetic algorithms. Computational Materials Science, 46(4), 821–827. https://doi.org/10.1016/j.commatsci.2009.04.023
Chakraborti, N. (2004). Genetic algorithms in materials design and processing. International Materials Reviews, 49(3–4), 246–260. https://doi.org/10.1179/095066004225021909
Chakraborti, N. (2023). Data-Driven Evolutionary Modeling in Materials Technology. CRC Press. https://doi.org/10.1201/9781003201045
Choudhary, K., & DeCost, B. (2021). Atomistic line graph neural network for improved materials property predictions. npj Computational Materials, 7, Article 185. https://doi.org/10.1038/s41524-021-00650-1
Choudhary, K., Garrity, K. F., Reid, A. C. E., DeCost, B., Biacchi, A. J., Hight Walker, A. R., Trautt, Z., Hattrick-Simpers, J., Gilad Kusne, A., Centrone, A., Davydov, A., Jiang, J., Pachter, R., Cheon, G., Reed, E., Agrawal, A., Qian, X., Sharma, V., Zhuang, H., … Tavazza, F. (2020). The joint automated repository for various integrated simulations (JARVIS) for data-driven materials design. npj Computational Materials, 6, Article 173. https://doi.org/10.1038/s41524-020-00440-1
Curtarolo, S., Hart, G. L. W., Buongiorno Nardelli, M., Mingo, N., Sanvito, S., & Levy, O. (2013). The high-throughput highway to computational materials design. Nature Materials, 12(3), 191–201. https://doi.org/10.1038/nmat3568
Dimitriu, R. C., Bhadeshia, H. K. D. H., Fillon, C., & Poloni, C. (2009). Strength of ferritic steels: Neural networks and genetic programming. Materials and Manufacturing Processes, 24(1), 10–15. https://doi.org/10.1080/10426910802539796
Elsken, T., Metzen, J. H., & Hutter, F. (2019). Neural architecture search: A survey. Journal of Machine Learning Research, 20(55), 1–21. https://doi.org/10.48550/arXiv.1808.05377
Giri, B. K., Pettersson, F., Saxén, H., & Chakraborti, N. (2013). Genetic programming evolved through bi-objective genetic algorithms applied to a blast furnace. Materials and Manufacturing Processes, 28(7), 776–782. https://doi.org/10.1080/10426914.2013.763953
Halder, C., Madej, L., Pietrzyk, M., & Chakraborti, N. (2015). Optimization of cellular automata model for the heating of dual-phase steel by genetic algorithm and genetic programming. Materials and Manufacturing Processes, 30(4), 552–562. https://doi.org/10.1080/10426914.2014.994765
Hamilton, W. L., Ying, R., & Leskovec, J. (2017). Inductive representation learning on large graphs. In I. Guyon, U. von Luxburg, S. Bengio, H. Wallach, R. Fergus, S. Vishwanathan, & R. Garnett (Eds.), Advances in Neural Information Processing Systems 30 (pp. 1024–1034). Curran Associates.
Hernandez, A., Balasubramanian, A., Yuan, F., Mason, S. A. M., & Mueller, T. (2019). Fast, accurate, and transferable many-body interatomic potentials by symbolic regression. npj Computational Materials, 5, Article 112. https://doi.org/10.1038/s41524-019-0249-1
Jain, A., Ong, S. P., Hautier, G., Chen, W., Richards, W. D., Dacek, S., Cholia, S., Gunter, D., Skinner, D., Ceder, G., & Persson, K. A. (2013). Commentary: The Materials Project: A materials genome approach to accelerating materials innovation. APL Materials, 1(1), Article 011002. https://doi.org/10.1063/1.4812323
Jennings, P. C., Lysgaard, S., Strabo Hummelshøj, J., Vegge, T., & Bligaard, T. (2019). Genetic algorithms for computational materials discovery accelerated by machine learning. npj Computational Materials, 5, Article 46. https://doi.org/10.1038/s41524-019-0181-4
Kipf, T. N., & Welling, M. (2017). Semi-supervised classification with graph convolutional networks. Openreview.net. https://openreview.net/forum?id=SJU4ayYgl
Klicpera, J., Groß, J., & Günnemann, S. (2020). Directional message passing for molecular graphs. Openreview.net. https://openreview.net/forum?id=WBCf7qTwcR
Krzywda, M. (2025). Graph neural network architecture search via hybrid genetic algorithm with parallel tempering. In M. Cha, C. Park, N. Park (Eds.), CIKM ‘25: Proceedings of the 34th ACM International Conference on Information and Knowledge Management (pp. 6793–6796). Association for Computing Machinery. https://doi.org/10.1145/3746252.3761661
Krzywda, M. (2026). CTGP – Concat-Tree Genetic Programming for GNN Architecture Search [Source code]. Github. https://github.com/mkrzywda/ctgt
Krzywda, M., Liu, Y., Łukasik, S., & Gandomi, A. H. (2025). Unveiling the search space of simple contrastive graph clustering with Cartesian genetic programming. In G. Ochoa (Ed.), GECCO ‘25 Companion: Proceedings of the Genetic and Evolutionary Computation Conference Companion (pp. 2380–2383). Association for Computing Machinery. https://doi.org/10.1145/3712255.3734538
Krzywda, M., Lukasik, S., & Gandomi, A. H. (2025a). Applying evolutionary techniques to enhance graph convolutional networks for node classification: case studies. In M. Bolanowski, M. Ganzha, L. Maciaszek, M. Paprzycki, D. Ślęzak (Eds.), Proceedings of the 20th Conference on Computer Science and Intelligence Systems (FedCSIS 2025) (pp. 321–326). PTI, IEEE. https://doi.org/10.15439/2025F0041
Krzywda, M., Łukasik, S., & Gandomi, A. H. (2025b). Linear genetic programming for design graph neural networks for node classification. In G. Ochoa, B. Filipič (Eds.), GECCO ’25 Companion: Proceedings of the Genetic and Evolutionary Computation Conference Companion (pp. 2167–2171). Association for Computing Machinery. https://doi.org/10.1145/3712255.3734278
Koza, J. R. (1992). Genetic Programming: On the Programming of Computers by Means of Natural Selection. MIT Press.
Mahanta, B. K., & Chakraborti, N. (2021). Evolutionary data driven modelling and many objective optimization of non linear noisy data in the blast furnace iron making process. Computer Methods in Materials Science, 21(3), 163–175. https://doi.org/10.7494/cmms.2021.3.0733
Miller, J. F. (2020). Cartesian genetic programming: its status and future. Genetic Programming and Evolvable Machines, 21(1–2), 129–168. https://doi.org/10.1007/s10710-019-09360-6
Roscher, R., Bohn, B., Duarte, M. F., & Garcke, J. (2020). Explainable machine learning for scientific insights and discoveries. IEEE Access, 8, 42200–42216. https://doi.org/10.1109/ACCESS.2020.2976199
Roy, S., Saini, B. S., Chakrabarti, D., & Chakraborti, N. (2020). Mechanical properties of micro-alloyed steels studied using a evolutionary deep neural network. Materials and Manufacturing Processes, 35(6), 611–624. https://doi.org/10.1080/10426914.2019.1660786
Schmidt, J., Marques, M. R. G., Botti, S., & Marques, M. A. L. (2019). Recent advances and applications of machine learning in solid-state materials science. npj Computational Materials, 5, 83. https://doi.org/10.1038/s41524-019-0221-0
Schütt, K. T., Sauceda, H. E., Kindermans, P.-J., Tkatchenko, A., & Müller, K.-R. (2018). SchNet – A deep learning architecture for molecules and materials. The Journal of Chemical Physics, 148(24), 241722. https://doi.org/10.1063/1.5019779
Suganuma, M., Shirakawa, S., & Nagao, T. (2017). A genetic programming approach to designing convolutional neural network architectures. In P. A. N. Bosman (Ed.), GECCO ‘17: Proceedings of the Genetic and Evolutionary Computation Conference (pp. 497–504). Association for Computing Machinery. https://doi.org/10.1145/3071178.3071229
Veličković, P., Cucurull, G., Casanova, A., Romero, A., Liò, P., & Bengio, Y. (2018). Graph attention networks. ArXiv. https://doi.org/10.48550/arXiv.1710.10903
Wu, F., Zhang, T., de Souza Jr., A. H., Fifty, C., Yu, T., & Weinberger, K. Q. (2019). Simplifying Graph Convolutional Networks. ArXiv. https://doi.org/10.48550/arXiv.1902.07153
Xie, T., & Grossman, J. C. (2018). Crystal graph convolutional neural networks for an accurate and interpretable prediction of material properties. Physical Review Letters, 120(14), Article 145301. https://doi.org/10.1103/PhysRevLett.120.145301
Xu, K., Hu, W., Leskovec, J., & Jegelka, S. (2019). How powerful are graph neural networks? ArXiv. https://doi.org/10.48550/arXiv.1810.00826
Yuan, F., & Mueller, T. (2017). Identifying models of dielectric breakdown strength from high-throughput data via genetic programming. Scientific Reports, 7, Article 17594. https://doi.org/10.1038/s41598-017-17535-3