SWSihong Wu

Sihong Wu

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Sihong Wu has a total of 31 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Demystifying and unifying ELBO for Probabilistic Deep Learning Inversion

    by Jiajia Sun, Sihong Wu

    Variational inference has emerged as a computationally efficient framework for solving Bayesian inverse problems. Rather than relying on random sampling in high-dimensional model spaces, it reformulates probabilistic inference as an optimization problem, marking a fundamental shift from sampling-based to optimization-based approaches. Its objective, the evidence lower bound (ELBO), is central to modern deep generative approaches, but its form can look different from one model class to another. This article derives the ELBOs step by step for two representative models used in probabilistic geophysical inversion: conditional variational autoencoders (cVAEs) and normalizing flows (NFs). The cVAE objective appears as a reconstruction term minus a latent-space Kullback–Leibler (KL) divergence, whereas the NF objective appears as data likelihood plus prior plus entropy. We show that these are not separate principles. Both are direct substitutions into the same generic ELBO identity. The apparent difference comes from the modeling choice of what counts as the observed variable and what is the hidden variable: in cVAE training, the hidden variable is the latent code; in unsupervised NF inversion, the hidden variable is the Earth model itself.

  • Probabilistic Inversion of Airborne Electromagnetic Data Using Conditional Variational Autoencoder: Imaging Hydrothermal Systems at Yellowstone

    by Chenyu Zhou, Sihong Wu, Jiajia Sun, Jiefu Chen, Yueqin Huang

    Airborne electromagnetic (AEM) inversion is inherently ill-posed, resulting in challenges for reliable subsurface interpretation. While Bayesian methods offer principled uncertainty quantification, their high computational cost limits scalability. We propose a conditional variational autoencoder (CVAE) framework for fast, probabilistic AEM inversion. By introducing a latent variable and conditioning on both AEM data and transceiver height, the CVAE learns the mapping relationship from latent space to posterior resistivity distributions. Synthetic tests demonstrate the CVAE’s noise robustness and its ability to reconstruct the model posterior. A comparison with McMC-based inversion is also discussed to highlight methodological and computational differences. Applied to AEM data acquired from Yellowstone region, the CVAE effectively recovers key hydrothermal features and provides quantified uncertainty estimates. This approach achieves orders-of-magnitude speed-up over traditional Bayesian methods, enabling large-scale, uncertainty-informed subsurface resistivity imaging.


  • Invertible neural network for real-time inversion and uncertainty quantification of ultra-deep resistivity measurements

    by George Bittar, Sihong Wu, Yawei Su, Shubin Zeng, Jiajia Sun, Xuqing Wu, Yueqin Huang, Jiefu Chen


  • Natural hydrogen exploration by joint sparse inversion of geophysical measurements and integrated geological interpretation

    by Yawei Su, Sihong Wu, Jiajia Sun, Xuqing Wu, Yueqin Huang, Jiefu Chen, Ligang Lu, Xiaolong Wei, Rodolfo Christiansen


  • Characterizing Marine Magnetic Anomalies: A Machine Learning Approach to Advancing the Understanding of Oceanic Crust Formation

    by S. Wu, S. Thoram, J. Sun, W. W. Sager, J. Chen

    Linear magnetic anomalies (LMA), resulting from Earth's magnetic field reversals recorded by seafloor spreading serve as crucial evidence for oceanic crust formation and plate tectonics. Traditionally, LMA analysis relies on visual inspection and manual interpretation, which can be subject to biases due to the complexities of the tectonic history, uneven data coverage, and strong local anomalies associated with seamounts and fracture zones. In this study, we present a Machine learning (ML)‐based framework to identify LMA, determine their orientations and distinguish spatial patterns across oceans. The framework consists of three stages and is semi‐automated, scalable and unbiased. First, a generation network produces artificial yet realistic magnetic anomalies based on user‐specified conditions of linearity and orientation, addressing the scarcity of the labeled training dataset for supervised ML approaches. Second, a characterization network is trained on these generated magnetic anomalies to identify LMA and their orientations. Third, the detected LMA features are clustered into groups based on predicted orientations, revealing underlying spatial patterns, which are directly related to propagating ridges and tectonic activity. The application of this framework to magnetic data from seven areas in the Atlantic and Pacific oceans aligns well with established magnetic lineations and geological features, such as the Mid‐Atlantic Ridge, Reykjanes Ridge, Galapagos Spreading Center, Shatsky Rise, Juan de Fuca Ridge and even Easter Microplate and Galapagos hotspot. The proposed framework establishes a solid foundation for future data‐driven marine magnetic analyses and facilitates objective and quantitative geological interpretation, thus offering the potential to enhance our understanding of oceanic crust formation.


  • Frontiers in electromagnetic geophysics — Introduction

    by Sihong Wu, Sizhuang Deng, Xiaolong Wei, Muhammad Rizwan Asif, Giulio Vignoli, Colin G. Farquharson


  • Open Access

    Real-time dual-parameter full-waveform inversion of GPR data based on robust deep learning

    by Jiyan Xue, Qinghua Huang, Sihong Wu, Li Zhao, Bowen Ma

    SUMMARY: Ground penetrating radar (GPR) is becoming an increasingly important tool for understanding the shallow electrical structures of the Earth and planets due to its adaptability to harsh detection environments, efficient data acquisition and accurate detection results. GPR full-waveform can simultaneously constrain the permittivity and resistivity of the medium, providing more comprehensive geophysical information and reducing the non-uniqueness of inversion. However, given the highly non-linear inverse problem and the massive data resulted from high temporal and spatial samplings, traditional full-waveform inversion algorithms are prohibitively costly. Inspired by Google's vision semantic segmentation system, we develop a robust deep learning-guided network that integrates geology and geophysics knowledge to support the real-time translation of zero-offset GPR data into dual-parameter electrical structures. We test our proposed network using synthetic data, which demonstrates that the algorithm can provide an accurate dual-parameter electrical model from a GPR sounding in milliseconds on a common laptop PC, exhibiting high robustness and adaptability to noise interference and extreme values of model parameters. We also apply our network to field data gathered for pollutant investigation in the United States. The resulting dual-parameter structure provides a more comprehensive and realistic depiction of subsurface electrical properties and reveals the migration and ageing of pollutants. Our algorithm's real-time and accurate advantages are expected to further unleash the potential of GPR technology and enable it to play a more significant role in earth and planetary exploration.


  • Open Access

    Physics-guided deep learning-based inversion for airborne electromagnetic data

    by Sihong Wu, Qinghua Huang, Li Zhao

    SUMMARY: The Earth's subsurface structure provides critical insights into sustainable resource management and geologic evolution. The airborne electromagnetic (AEM) method is an efficient data acquisition technique and can be used to image the underground resistivity structure with high spatial resolution. However, inversion of the increasingly huge volume of AEM data poses a heavy computational burden. In this study, we develop a hybrid deep learning-based approach by using the physics-guided neural network (PGNN) which incorporates the governing physical laws into the loss function to solve the AEM inverse problem. The PGNN integrates the strength of data-driven method for representation learning with electromagnetic laws and allows for the underlying physical constraints to be strictly satisfied. We validate the effectiveness of our approach using both synthetic and field datasets. Compared with the classic Gauss–Newton method, our PGNN inversion system shows strong robustness against multiple noise sources and reduces the risk of being trapped in local extrema. Moreover, the PGNN-inverted results are physically more consistent with the AEM observations compared to the purely data-driven approach. Application to the field AEM data from Northern Australia demonstrates that the PGNN-based inversion framework effectively estimates the subsurface electrical properties with considerable lateral continuity and significantly higher efficiency, completing the inversion of more than 2734000 AEM soundings taking only minutes on a common PC. Our proposed PGNN-based method shows great promise for large-scale underground resistivity imaging, and the well-identified subsurface resistivity structure can effectively improve our understanding of resource distributions and geological hazards.


  • Detection of ULF Geomagnetic Anomalies Prior to the Tohoku-Oki Earthquake by the Multireference Station Method

    by Jiyan Xue, Qinghua Huang, Sihong Wu, Li Zhao

    The ultra-low-frequency (ULF) electromagnetic anomaly has been considered one of the earthquake precursory signals with the potential for short-term prediction. As such, effectively detecting ULF anomalies is important for mitigating earthquake disasters. Given the comprehensive coverage of geomagnetic networks in Japan, the 2011 Tohoku-Oki earthquake (M9.0) provided a tremendous opportunity for investigating the characteristics and mechanisms of ULF anomalies. Previous studies reporting detections of ULF anomalies triggered by the Tohoku-Oki earthquake have been questioned on the grounds of insufficient reliability due to technological limitations. In this study, we employ a multi-reference station data-quality-weighted method to detect the ULF anomaly. Comparison with traditional single-reference station method demonstrates the robustness of our detection results. Statistical test also indicates that the ULF anomaly appearing approximately two months before the earthquake was driven by physical processes. Geomagnetic storm analysis further rules solar activity as the cause of the anomaly. Spatial distribution of the anomaly amplitude reveals a decrease in the anomaly energy with increasing epicentral distance, implying a strong association between the anomaly and the earthquake. Our study is helpful for understanding the possible connection between the ULF electromagnetic precursors and the seismogenic processes of strong earthquakes, which contributes to achieving the ultimate goal of short-term earthquake prediction.


  • Open Access

    Near real-time subsurface structure imaging using airborne electromagnetic data based on deep learning

    by Sihong Wu, Qinghua Huang, Li Zhao

    The airborne electromagnetic (AEM) method is a modern technique in geophysical surveys with the merits of terrain adaptability and acquisition efficiency, and can image the electrical structure of the Earth’s subsurface down to several hundred meters. AEM survey has been applied extensively in mineral exploration, groundwater monitoring, environment investigation and geological mapping. However, the expansion of survey area and spatial sampling bring about huge volumes of AEM data, which present a serious computational challenge for rapid AEM inversion. Inspired by Google’s neural machine translation system, we develop a fast inversion system guided by deep learning to translate AEM data directly into subsurface resistivity structures. Synthetic tests demonstrate that our proposed inversion system has strong noise robustness and can provide reliable inversion results much more efficiently than the conventional Gauss-Newton algorithm. In the field data application, our system obtains robust inversion results of more than 740,000 AEM soundings acquired by the U.S. Geological Survey from the Leach Lake Basin in California in seconds on a common PC. The inverted images coincide exactly with previous studies of the local geological environment and clearly outline the geometries of the lake, faults and surrounding mountains. Our inversion system can support near real-time underground subsurface imaging for AEM surveys and inject new vigor into resource exploration and tectonic studies.