Why Diffusion Models Fail for Inverse Materials Design: Mode Collapse, Compositional Violation, and Stability Loss

iamrp.net-gpjC-1791491021-m112777376 1791491021 Institute for Advanced Materials Research Press info@iamrp.net Institute for Advanced Materials Research Press Journal of Computational and Data-Driven Materials Engineering J. Comput. Data-Driven Mater. Eng. 3149-9368 2025 4 1 Why Diffusion Models Fail for Inverse Materials Design: Mode Collapse, Compositional Violation, and Stability Loss Hassan Ali Mariam Farooq Usman Shah 2025 10.68159/m112777376 https://iamrp.net/pub/journal/2/article/m112777376 Ho J, Jain A, Abbeel P. Denoising diffusion probabilistic models. In: Advances in neural information processing systems. 2020;33:6840-51. Corvi R, Cozzolino D, Poggi G, Nagano K, Verdoliva L. Intriguing properties of synthetic images: From generative adversarial networks to diffusion models. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops; 2023. p. 973-82. Song Z, Meng Z, King I. A diffusion-based pre-training framework for crystal property prediction. Proc AAAI Conf Artif Intell. 2024;38(8):8993-9001. 10.1609/aaai.v38i8.28748 Chen L, Zhang W, Nie Z, Li S, Pan F. Generative models for inverse design of inorganic solid materials. J Mater Inf. 2021;1(1):4. 10.20517/jmi.2021.07 De Breuck PP, Wang HC, Rignanese GM, Botti S, Marques MAL. Generative AI for crystal structures: A review. npj Comput Mater. 2025;11(1):370. 10.1038/s41524-025-01881-2 Song Y, Siriwardane EMD, Zhao Y, Hu J. Computational discovery of new 2D materials using deep learning generative models. ACS Appl Mater Interfaces. 2021;13(45):53303-13. 10.1021/acsami.1c01044 Zeni C, Pinsler R, Zügner D, Fowler A, Horton M, Fu X, et al. A generative model for inorganic materials design. Nature. 2025;639(8055):624-32. 10.1038/s41586-025-08628-5 Luo X, Wang Z, Wang Q, Shao X, Lv J, Wang L, et al. CrystalFlow: A flow-based generative model for crystalline materials. Nat Commun. 2025;16(1):9267. 10.1038/s41467-025-64364-4 Park H, Onwuli A, Walsh A. Exploration of crystal chemical space using text-guided generative artificial intelligence. Nat Commun. 2025;16(1):4379. 10.1038/s41467-025-59636-y Manduchi L, Pandey K, Bamler R, Cotterell R, Däubener S, Fellenz S, et al. On the challenges and opportunities in generative AI. arXiv Preprint. 2024. 10.48550/arXiv.2403.00025 Ruthotto L, Haber E. An introduction to deep generative modeling. GAMM Mitt. 2021;44(2):e202100008. 10.1002/gamm.202100008 Barsha FL, Eberle W. An in-depth review and analysis of mode collapse in generative adversarial networks. Mach Learn. 2025;114(6):141. 10.1007/s10994-025-06772-7 Long T, Fortunato NM, Opahle I, Zhang Y, Samathrakis I, Shen C, et al. Constrained crystals deep convolutional generative adversarial network for the inverse design of crystal structures. npj Comput Mater. 2021;7(1):66. 10.1038/s41524-021-00526-4 Lalena JN, Cleary DA, Hardouin Duparc OBM. Principles of inorganic materials design. 3rd ed. Hoboken (NJ): Wiley; 2020. 720 p. Osting B, Wang D. A diffusion generated method for orthogonal matrix-valued fields. Math Comput. 2020;89(322):515-50. 10.1090/mcom/3473 Alverson M, Baird SG, Murdock R, Ho ESH, Johnson J, Sparks TD. Generative adversarial networks and diffusion models in material discovery. Digit Discov. 2024;3(1):62-80. 10.1039/D3DD00137G Schubert US, Hüsing N. Synthesis of inorganic materials. 4th ed. Weinheim: Wiley-VCH; 2019. 424 p. Zhu R, Tian SIP, Ren Z, Li J, Buonassisi T, Hippalgaonkar K. Predicting synthesizability using machine learning on databases of existing inorganic materials. ACS Omega. 2023;8(9):8210-8. 10.1021/acsomega.2c04856 Zhao Y, Siriwardane EMD, Wu Z, Fu N, Al-Fahdi M, Hu M, et al. Physics guided deep learning for generative design of crystal materials with symmetry constraints. npj Comput Mater. 2023;9(1):38. 10.1038/s41524-023-00987-9 Lyngby P, Thygesen KS. Data-driven discovery of 2D materials by deep generative models. npj Comput Mater. 2022;8(1):232. 10.1038/s41524-022-00923-3 Cui H, Krzakala F, Vanden-Eijnden E, Zdeborová L. Analysis of learning a flow-based generative model from limited sample complexity. arXiv Preprint. 2023. 10.48550/arXiv.2310.03575 Goodfellow I, Pouget-Abadie J, Mirza M, Xu B, Warde-Farley D, Ozair S, et al. Generative adversarial networks. Commun ACM. 2020;63(11):139-44. 10.1145/3422622 Creswell A, White T, Dumoulin V, Arulkumaran K, Sengupta B, Bharath AA. Generative adversarial networks: An overview. IEEE Signal Process Mag. 2018;35(1):53-65. 10.1109/MSP.2017.2765202 Chen Z, Meng Z, He T, Li H, Cao J, Xu L, et al. Crystal structure prediction meets artificial intelligence. J Phys Chem Lett. 2025;16(10):2581-91. 10.1021/acs.jpclett.4c03727 Hu Z, Yang Z, Salakhutdinov R, Xing EP. On unifying deep generative models. arXiv Preprint. 2017. 10.48550/arXiv.1706.00550 Jahanian A, Chai L, Isola P. On the "steerability" of generative adversarial networks. arXiv Preprint. 2019. 10.48550/arXiv.1907.07171 Shin I, Yang C, Chen LC. Deeply supervised flow-based generative models. In: Proceedings of the IEEE/CVF International Conference on Computer Vision; 2025. p. 16535-44. 10.1109/ICCV51701.2025.01535