Oral: Inverse Problem Techniques for Pattern Recognition in HR-TEM Images of Graphene Oxide

Graphene oxide (GO) is widely used in flexible electronics and nanotechnology. High-resolution transmission electron microscopy (HR-TEM) reveals atomic-level structures; however, interpreting these images is challenging owing to noise and polycrystalline complexities. This study explores inverse problem techniques, specifically Sparse Identification of Nonlinear Dynamics (SINDy), to enhance pattern recognition in HR-TEM images of GO and identify key structural features, such as hydroxyl and epoxy bridges. By applying SINDy to the simulated and experimental data, it was inferred a set of potential governing equations that capture these patterns, even in noisy datasets. This approach aims to explain GO's structural properties of GO and to advance its application in nanotechnology.