In the past few years machine learning techniques have increasingly been employed for solving problems in scientific computing, i.e., the approximation of differential equations. Such efforts are now commonly referred to as “scientific machine learning” [1]. Just as with classical numerical methods, it has been observed that scientific machine learning benefits from preserving structural properties of the differential equations, e.g., symplecticity [2] and symmetries [3]. This minisymposium invites speakers to present work on preserving geometric structure in machine learning models for applications like system discovery from experimental [4] data and reduced order modeling [5, 6]. Presentations can encompass novel algorithms as well as software implementations and comparisons of existing approaches. REFERENCES [1] Baker N, et al. Workshop report on basic research needs for scientific machine learning: Core technologies for artificial intelligence. USDOE Office of Science (SC), Washington, DC (United States), 2019. [2] Jin P, et al. SympNets: Intrinsic structure-preserving symplectic networks for identifying Hamiltonian systems. Elsevier Neural Networks 132, 2020. [3] Lishkova Y, et al. Discrete Lagrangian neural networks with automatic symmetry discovery. Elsevier IFAC-PapersOnLine, 2023. [4] Greydanus S, Dzamba M, Yosinski J. Hamiltonian neural networks. Advances in neural information processing systems 32, 2019. [5] Buchfink P, Glas S, and Haasdonk B. Symplectic model reduction of Hamiltonian systems on nonlinear manifolds and approximation with weakly symplectic autoencoder. SIAM Journal on Scientific Computing 45.2, 2023. [6] Brantner B, Kraus M. Symplectic autoencoders for model reduction of Hamiltonian systems. arXiv preprint arXiv:2312.10004, 2023.