I am a fifth-year Ph.D. candidate in Civil Engineering at Tongji University, China, advised by Prof.Fengshou Zhang. I am currently a research scholar in the Sibley School of Mechanical and Aerospace Engineering at Cornell University, working with Prof.Jian-Xun Wang. I have been actively collaborating with leading researchers, including Prof.Xianda Shen from Tongji University, and Prof.Bicheng Yan from King Abdullah University of Science and Technology (KAUST).

My research lies at the intersection of artificial intelligence and geoscience, advancing the energy transition to a sustainable, carbon-neutral future. I am interested in developing AI-enabled computational frameworks for subsurface energy systems, with applications to geological carbon sequestration, shale gas exploitation, and beyond.

📝 Publications

2026

From simulator to surrogate: GANs for spatiotemporal modeling of subsurface flow in porous media. Zeeshan Tariq*, Zhao Feng, Bicheng Yan, Hussein Hoteit. Computational Geosciences.

2025

ArXiv

Generative Latent Diffusion Model for Inverse Modeling and Uncertainty Analysis in Geological Carbon Sequestration

Zhao Feng, Xin-Yang Liu, Meet Hemant Parikh, Junyi Guo, Pan Du, Bicheng Yan*, Jian-Xun Wang*

Code:

Our proposed CoNFiLD-geo combines conditional neural field encoding with Bayesian conditional latent-space diffusion, enabling zero-shot conditional generation of geomodels and reservoir responses across complex geometries and grid structures.
The model is pretrained unconditionally in a self-supervised manner, followed by a Bayesian posterior sampling process, allowing for data assimilation for unseen states without task-specific retraining.

Journal of Hydrology

SURGIN: SURrogate-guided Generative INversion for subsurface multiphase flow with quantified uncertainty

Zhao Feng, Bicheng Yan, Luanxiao Zhao, Xianda Shen, Renyu Zhao, Wenhao Wang, Fengshou Zhang*

Code:

SURGIN integrates a U-Net enhanced Fourier Neural Operator (U-FNO) surrogate with a score-based generative model (SGM), framing the conditional generation as a surrogate prediction-guidance process in a Bayesian perspective.
An unconditional SGM is first pretrained in a self-supervised manner to capture the geological prior, after which posterior sampling is performed by leveraging a differentiable U-FNO surrogate to enable efficient forward evaluations conditioned on unseen observations.

Advances in Water Resources

A hybrid CNN-transformer surrogate model for the multi-objective robust optimization of geological carbon sequestration

Zhao Feng, Bicheng Yan, Xianda Shen, Fengshou Zhang*, Zeeshan Tariq, Weiquan Ouyang, Zhilei Han

Code:

A hybrid CNN-Transformer surrogate model is developed to simulate time-varying CO₂ injection into 3D heterogeneous geological fields.
Multi-objective robust optimization is implemented by integrating CNN-Transformer with Non-dominated Sorting-based genetic algorithm II.

CoSwinNet: a conditional Swin Transformer multimodal surrogate model for subsurface multiphase flow. Zhao Feng, Zeeshan Tariq*, Zhong Zhang, Peilin Zhao, Renyu Zhao, Wenhao Wang, Xinwo Huang, Bicheng Yan, Xianda Shen*, Fengshou Zhang. Fuel.
CO₂ sequestration and mineralization in basalts: Insights from a deep learning-based surrogate model. Weiquan Ouyang, Zhao Feng, Fengshou Zhang*, Zhao Xia, Xianda Shen. Engineering Geology.

2024

Gas Science and Engineering

An encoder-decoder ConvLSTM surrogate model for simulating geological CO₂ sequestration with dynamic well controls

Zhao Feng, Zeeshan Tariq, Xianda Shen*, Bicheng Yan*, Xuhai Tang, Fengshou Zhang

Code:

An Encoder-Decoder Convolutional LSTM (ED-ConvLSTM) is developed for simulating CO₂ flow under dynamic injection rates.
The proposed network has excellent capability in extrapolating to unseen time steps and to different geological settings.

TransUNet: Accelerating Multiphase Flow Prediction in Geological Carbon Storage Systems. Zeeshan Tariq, Zhao Feng, Hussein Hoteit, Moataz Abualsaud, Xupeng He, Muhammad AlMajid, Shuyu Sun, Bicheng Yan. Abu Dhabi International Petroleum Exhibition and Conference.
Transformer-Based Surrogate Model for the Optimization of Geological Carbon Sequestration. Zhao Feng, Weiquan Ouyang, Zeeshan Tariq, Zhilei Han, Xianda Shen, Bicheng Yan, Fengshou Zhang. Asian Rock Mechanics Symposium.

2023

ISRM Congress

A Data-Driven Model for the Prediction of Stimulated Reservoir Volume (SRV) Evolution During Hydraulic Fracturing

Zhao Feng, Fengshou Zhang, Xianda Shen, Bichen Yan, Zhaowei Chen

Code: | Slides: [slides]

A data-driven model is developed to use field monitored hydraulic fracturing parameters and microseismicity to predict the stimulated reservoir volume (SRV) ahead of time.
We use a voxelized method to calculate the SRV and a LSTM network to handle the time sequence regression task.

📖 Education

2021.09 - Now, Ph.D. in Civil Engineering, Tongji University.
2017.09 - 2021.06, B.S. in Mining Engineering, China University of Mining and Technology.

💼 Work Experience

2025.08 - Now, Research Scholar, Cornell University, USA.
2025.01 - 2025.07, Research Scholar, University of Notre Dame, USA.
2023.09 - 2023.12, Research Scholar, King Abdullah University of Science and Technology, Saudi Arabia.
2019.02 - 2019.06, Visiting Student, Technical University of Leoben, Austria.

🎖 Honors and Awards

2026 Best Paper Award, American Rock Mechanics Association.
2025 Lingjun Talent Fellowship, Tongji University.
2021 Outstanding Undergraduate, China University of Mining and Technology.
2020 Zhang Guangshi Scholarship, China University of Mining and Technology.
2017,2019 First-class Scholarship, China University of Mining and Technology.
2018 Chinese Government Scholarship, China Scholarship Council.

💬 Talks and Posters

2026.6, “Learning spatiotemporal dynamics using pretrained large language foundation models”, USNC-TAM @ Pasadena. [Talk]
2026.6, “Generative Latent Diffusion Model for Inverse Modeling and Uncertainty Analysis in Geological Carbon Sequestration”, MIT ERL Meeting @ Cambridge. [Poster]
2025.12, “SURGIN: SURrogate–guided Generative INversion for Geological Carbon Sequestration”, AGU Annual Meeting @ New Orleans. [Poster].
2024.06, “An encoder-decoder ConvLSTM surrogate model for simulating geological CO₂ sequestration with dynamic well controls”, International Carbon Capture, Utilization and Storage (CCUS) Conference @ Beijing. [Talk].
2023.11, “A deep-learning-based Convolutional-LSTM surrogate model for simulating geological CO₂ sequestration with dynamic well controls”, International Geomechanics Symposium @ Al Khobar. [Talk]