https://scholarworks.bwise.kr/hanyang/handle/2021.sw.hanyang/154 2026-07-03T23:11:04Z 2026-07-03T23:11:04Z Nguyen, Le-Anh-Nhat Lee, Yong-Gook Park, Duhee Tsai, Chi-Chin https://scholarworks.bwise.kr/hanyang/handle/2021.sw.hanyang/212945 2026-06-02T05:00:16Z 2026-12-01T00:00:00Z

Title: Development of machine learning-based site amplification models for Japan from borehole recordings Authors: Nguyen, Le-Anh-Nhat; Lee, Yong-Gook; Park, Duhee; Tsai, Chi-Chin Abstract: A large number of site amplification models have been developed using regression and machine learning (ML) approaches. Although ML models generally outperform traditional methods in predicting site amplification, the effects of specific site and motion parameters on model accuracy remain insufficiently explored. Using a meta-dataset of earthquake recordings from Japan’s Kiban Kyoshin Network (KiK-net), we trained six ML-based site amplification models: random forest (RF), extreme gradient boosting (XGB), and deep neural network (DNN), together with their hybrid variants incorporating Bayesian optimization (BO). A sensitivity analysis using RF evaluated how combinations of input proxies influence predictive performance, leading to the identification of an optimal proxy configuration. Among the six models trained with this configuration, BO-DNN performed best at period T < 0.1 s, whereas BO-XGB showed superior performance at T > 0.1 s. Shapley Additive exPlanations (SHAP) was used to rank proxy importance, identifying the peak frequency of the horizontal-to-vertical spectral-ratio curve (fp), the time-averaged shear-wave velocity up to 30 m (Vs30), borehole depth (BD), and borehole spectral acceleration averaged over 0.1–0.3 s (SS) as the most influential proxies. The proposed models demonstrate superior performance compared with two previously published models that were also developed using KiK-net data.

2026-12-01T00:00:00Z Chen, Mengdie Mangalathu, Sujith Jeon, Jong-Su https://scholarworks.bwise.kr/hanyang/handle/2021.sw.hanyang/217780 2026-07-02T02:00:16Z 2026-10-01T00:00:00Z

Title: Physics-guided residual learning framework for aftershock time history prediction using mainshock acceleration data Authors: Chen, Mengdie; Mangalathu, Sujith; Jeon, Jong-Su Abstract: Accurate aftershock acceleration time history prediction is essential for post-earthquake structural evaluation, particularly under data-scarce conditions. A physics-guided residual learning framework is developed to generate aftershock acceleration time histories from mainshock acceleration time histories and static site parameters. The proposed approach begins with a pseudo-aftershock acceleration time history constructed by applying exponential attenuation to the mainshock, which serves as a physically inspired prior for residual learning. Multiscale convolutional encoders extract temporal features, whereas static variables such as magnitude and site conditions are fused through gated embedding. A residual correction module guided by cross-attention refines the pseudo-aftershock to match the observed aftershock responses. A hybrid loss function ensures consistency in both time and frequency domains. Despite the limited number of training samples, the proposed residual learning model achieved consistently low prediction errors and showed promising predictive performance under a limited-event setting based on 140 paired sequences from 10 seismic events, including 20 test sequences. The architecture further enabled interpretable forecasting, revealing the manner in which mainshock dynamics and static attributes jointly influenced residual adjustments. On an independent test set, the model achieved a mean absolute error of 0.00553 g (where g denotes the gravitational acceleration), a root mean square error of 0.01304 g, and a coefficient of determination of 0.612. These findings highlight the feasibility of residual-based physically guided forecasting for realistic structural demand assessments.

2026-10-01T00:00:00Z Kim, Jinwoo https://scholarworks.bwise.kr/hanyang/handle/2021.sw.hanyang/212259 2026-04-21T00:00:16Z 2026-09-01T00:00:00Z

Title: Synergistic in-domain and out-of-domain learning to strengthen visual scene understanding in data-scarce, imbalanced construction settings Authors: Kim, Jinwoo Abstract: A scarcity of in-domain training images from target workplaces has hindered the broader adoption of visual scene understanding in construction. While out-of-domain images from non-target workplaces hold promising potential as supplementary training data, traditional approaches—put-it-all-together training and sequential finetuning—struggle with either over-generalization or catastrophic forgetting when both datasets are mixed due to distributional inconsistencies. To overcome this limitation, this article introduces a synergistic learning strategy that acquires domain-invariant visual knowledge from out-of-domain datasets while reinforcing in-domain predictive capabilities. Results show that the synergistic strategy consistently outperforms traditional approaches across various evaluation criteria, regardless of dataset size, imbalance, and distribution. Remarkably, the strategy achieves comparable or superior performance with merely half or one-twentieth of the in-domain dataset, showing robustness to diverse and challenging site conditions. Further studies further validate its effectiveness when tested on different object category and model architecture, as well as with an irrelevant out-of-domain dataset. These results strongly indicate that synergistic learning can effectively balance in-domain and out-of-domain visual knowledge, enhancing model’s performance and generalizability in diverse field conditions. These findings will be a solid foundation for advancing visual scene understanding in data-scarce, imbalanced industry settings including construction

2026-09-01T00:00:00Z He, Gaoliang Maitlo, Hubdar Ali Kim, Ki-Hyun https://scholarworks.bwise.kr/hanyang/handle/2021.sw.hanyang/213307 2026-06-17T00:00:15Z 2026-09-01T00:00:00Z

Title: From charge separation to hole confinement: A strategy for maximizing oxidative power in an n-n S-scheme using TiO2–Bi2S3 as a model photocatalyst Authors: He, Gaoliang; Maitlo, Hubdar Ali; Kim, Ki-Hyun Abstract: A novel strategy is developed to maximize the photocatalytic oxidative power of a composite by transitioning from conventional charge-separation models to a targeted hole-confinement mechanism using an n-n TiO2/Bi2S3 S-scheme heterojunction. It is demonstrated how high-potential oxidative species (ROS) are selectively sequestered through engineering of the internal electric field and interfacial band bending. The S-scheme pathway is explicitly confirmed by advanced operando analysis validation using in situ Kelvin probe force microscopy and irradiated X-ray photoelectron spectroscopy, revealing that high-potential holes accumulate in TiO2. Furthermore, the excited-state lifetime of photocarriers is shown by femtosecond transient absorption spectroscopy to nearly double for the TBS-10 heterojunction (898 ps) compared to TiO2 (449 ps). This significant extension indicates that charge carrier recombination is effectively suppressed. The optimized TBS-10 exhibits excellent photocatalytic degradation performance, achieving 100% formaldehyde removal efficiency, an apparent quantum yield of 0.077%, and a clean air delivery rate of 20.13 L min−1. The in-situ DRIFTS reveals a photocatalytic oxidation pathway proceeding through dioxymethylene and formate intermediate toward near-complete mineralization to CO2 and H2O. This work offers a mechanistically guided framework for advancing S-scheme architectures toward high-performance environmental remediation.

2026-09-01T00:00:00Z