AI has greatly improved epidemiological modeling, especially with the use of Physics-Informed Neural Networks (PINNs). Recently, Kolmogorov-Arnold Networks (KANs) have emerged, offering better interpretability and requiring fewer parameters. In this study, we apply KANs to compartmental models in epidemiology, focusing on multi-scale outbreaks, nonlinear infection dynamics, and diverse population distributions. We create Kolmogorov-Arnold-Informed Neural Networks (KINNs) and compare their performance with PINNs. Our results show that KINNs provide higher prediction accuracy and faster convergence in most cases. Additionally, KINNs excel in estimating model parameters, which is crucial for accurate disease modeling. This research highlights the potential of KANs to improve the precision and efficiency of epidemiological predictions, making them useful tools in managing infectious diseases.
@inproceedings{erland24,author={Ramadhan, Erland Rachmad},title={Kolmogorov-Arnold Informed Neural Networks: Advancing AI in Epidemiological Modeling},booktitle={International Symposium on Biomathematics},organization={IBMS},month=jul,year={2024}}
2020
Inorg. Chem. Front.
Octahedral morphology of NiO with (111) facet synthesized from the transformation of NiOHCl for the NOx detection and degradation: experiment and DFT calculation
Angga Hermawan, Adie Tri Hanindriyo, Erland Rachmad Ramadhan, and 6 more authors
Nitrogen oxides (NOx) are poisonous gases to humans and environment, and need to be monitored at an early stage. Accordingly, the design of facets on metal oxide semiconductors is an efficient approach to boost their gas sensing and photocatalytic performances due to the desirable active sites. However, in a rock-type structured NiO, a highly polar (111) exposed facets cannot be easily exposed due to their unfavorable thermodynamics. Herein, we demonstrate the synthesis of NiO with a dominant (111) facet from the transformation of NiOHCl with a layered structure. Among the crystal facets, NiO-Octa (111) exhibited the best NOx gas sensing response (16.5%) to the 300 ppb level and deNOx photocatalytic ability of over 50% under UV irradiation. DFT calculations revealed that the abundance of Ni atoms in the clean (111) surface layer allow the favorable adsorption of N adatoms, forming the Ni–N bond. The charge transfer occurring from NiO to the NO orbital was proven to be the cause for bond weakening and stretching from 1.1692 Å to 1.2231 Å, leading to NOx molecular decomposition, consistent with experimental results.
@article{hermawan20,author={Hermawan, Angga and Hanindriyo, Adie Tri and Ramadhan, Erland Rachmad and Asakura, Yusuke and Hasegawa, Takuya and Hongo, Kenta and Inada, Miki and Maezono, Ryo and Yin, Shu},title={Octahedral morphology of NiO with (111) facet synthesized from the transformation of NiOHCl for the NOx detection and degradation: experiment and DFT calculation},journal={Inorganic Chemistry Frontiers},month=jul,year={2020},volume={7},issue={18},pages={3431-3442},publisher={The Royal Society of Chemistry},doi={10.1039/D0QI00682C},url={http://dx.doi.org/10.1039/D0QI00682C}}