SELECTED WORK
From theory to physical impact.
On the Computational Limits of AI4S-RL: A Unified ε-N Analysis
Quantifies the minimum computational cost required for an AI4S surrogate to support reliable RL, connecting surrogate accuracy, physical resolution, and policy quality in one probabilistic framework.
Value-Priced Uncertainty: A Family of PPO-Compatible Exploration Bonuses
Introduces VALU, a PPO exploration bonus that weighs transition uncertainty by its local effect on advantage estimates; evaluated across six MuJoCo continuous-control tasks.
Reinforcement learning–based adaptive strategies for climate change adaptation
Uses reinforcement learning to coordinate protection, accommodation, and retreat under deep climate uncertainty, reducing coastal adaptation costs by up to 77%.
Blind-Window Forecasting: Real-Time Benchmarking and Multimodal Reconstruction for Tropical Cyclones
Introduces a 1980–2023 benchmark for forecasting with delayed reanalysis and timely satellite observations; RECAST-TC reconstructs missing recent storm context for track and intensity prediction.
PhysTC: A Physics-Enhanced Dataset and Architecture for High-Precision Tropical Cyclone Forecasting
Presents a globally harmonized 1950–2023 tropical-cyclone dataset and PhysTCN, a physics-enhanced model for forecasting storm tracks and intensity from coarse observations.
Hurricane Ida's blackout–heatwave compound risk in a changing climate
Shows how climate change amplifies compound blackout–heatwave risk and identifies the dominant physical drivers of future exposure.
Tropical cyclone–blackout–heatwave compound hazard resilience in a changing climate
Develops a climate-aware power-system resilience model and identifies cost-effective infrastructure hardening strategies.
RainNet: A large-scale imagery dataset and benchmark for spatial precipitation downscaling
Introduces a large real-world precipitation dataset and benchmark for learning high-resolution rainfall fields.
Climate change exacerbates hurricane flood hazards along US Atlantic and Gulf Coasts in spatially varying patterns
Quantifies how sea-level rise and changing tropical-cyclone climatology reshape hurricane flood hazards across the US Atlantic and Gulf Coasts.
Solar and wind energy enhances drought resilience and groundwater sustainability
Shows how solar and wind deployment can improve drought resilience and groundwater sustainability by easing water–food–energy trade-offs.
FaceProbe: Recovering HDR Environment Map via Masked Diffusion and Physical Preference
Recovers HDR environment maps from portraits through masked diffusion and relighting-based preference optimization, improving portrait relighting accuracy.
FastAvatar: Towards Unified and Fast 3D Avatar Reconstruction with Large Gaussian Reconstruction Transformers
A unified feedforward model that reconstructs high-fidelity 3D Gaussian avatars within seconds from a single image, multi-view observations, or monocular video.