Jinpai Zhao

Jinpai Zhao, Albert Cerrone, Joannes Westerink et al.

Single-agent systems (SAS) have become the default pattern for LLM-driven scientific workflows, but routing planning, tool use, and synthesis through a single context window comes with a well-known cost: as tool specifications and observational traces accumulate, the effective context available for each decision shrinks, and end-to-end reliability suffers. We present a multi-agent system (MAS) prototype for hydrodynamics in which specialized agents are coordinated through a Layer Execution Graph (LEG). A planner agent constructs query-specific execution topologies from natural-language routing heuristics that capture domain knowledge without hard-coding it as rigid control logic; specialist agents operate under strict tool allowlists and occupy complementary data-class roles. Between layers, consolidator agents fuse parallel outputs into concise briefs, and a reporter agent synthesizes the final response, while the runtime logs provenance for every tool invocation to support auditability. All benchmarks, ablations, and stress tests use Claude Sonnet~4.6 as the backbone model for both specialist and general-purpose agents. Evaluated on 37 queries spanning six complexity categories, the prototype achieves 93.6% factual precision with a 100% pass rate. Accuracy remains above 90% across runs from single-threaded to five independent parallel tracks, and under simulated loss of individual data sources the system degrades gracefully, still returning substantive partial answers. Together, these results suggest that planner-guided, graph-structured multi-agent orchestration can meaningfully alleviate the context-saturation bottlenecks that constrain monolithic single-agent architectures.

Benjamin Pachev, Prateek Arora, Jinpai Zhao et al.

Storm surge is a significant threat to coastal communities across the globe, responsible for loss of life and enormous property damage. Consequently, significant efforts have been expended to develop high-fidelity physics-based models for storm surge prediction. However, such models are often extremely computationally expensive and require supercomputing resources. In recent years, there has been a growing trend towards data-driven surrogate models, which approximate the capabilities of high-fidelity models at a tiny fraction of the computational cost. Most datasets of high-fidelity storm surge model output are limited to narrow geographical regions, with the majority focused on the continental United States and China. This trend is reflected in the scope of existing storm surge surrogate models. In this work, we present a novel dataset for training storm surge surrogate models with global applicability. The dataset consists of high-resolution peak surge output from the ADvanced CIRCulation (ADCIRC) model for over 15,000 landfalling synthetic storms distributed across the world. To the author's knowledge, it is the largest dataset of its kind ever assembled, and is unique in its global scope. We additionally present a machine learning model for peak storm surge based on computer vision architecture. The model is trained on our new global dataset and can accurately predict maximum storm surge in disparate geographical regions - including those for which few or no surrogate models exist. Both the dataset and accompanying model are publicly available, with the aim to support the development of additional storm surge models with global reach.

Jinpai Zhao, Nishant Panda, Yen Ting Lin et al.

We introduce HyCOP, a modular framework that learns parametric PDE solution operators by composing simple modules (advection, diffusion, learned closures, boundary handling) in a query-conditioned way. Rather than learning a monolithic map, HyCOP learns a policy over short programs - which module to apply and for how long - conditioned on regime features and state statistics. Modules may be numerical sub-solvers or learned components, enabling hybrid surrogates evaluated at arbitrary query times without autoregressive rollout. Across diverse PDE benchmarks, HyCOP produces interpretable programs, delivers order-of-magnitude OOD improvements over monolithic neural operators, and supports modular transfer through dictionary updates (e.g., boundary swaps, residual enrichment). Our theory characterizes expressivity and gives an error decomposition that separates composition error from module error and doubles as a process-level diagnostic.

Jinpai Zhao, Albert Cerrone, Eirik Valseth et al.

Storm surge forecasting plays a crucial role in coastal disaster preparedness, yet existing machine learning approaches often suffer from limited spatial resolution, reliance on coastal station data, and poor generalization. Moreover, many prior models operate directly on unstructured spatial data, making them incompatible with modern deep learning architectures. In this work, we introduce a novel approach that projects unstructured water elevation fields onto structured Red Green Blue (RGB)-encoded image representations, enabling the application of Convolutional Long Short Term Memory (ConvLSTM) networks for end-to-end spatiotemporal surge forecasting. Our model further integrates ground-truth wind fields as dynamic conditioning signals and topo-bathymetry as a static input, capturing physically meaningful drivers of surge evolution. Evaluated on a large-scale dataset of synthetic storms in the Gulf of Mexico, our method demonstrates robust 48-hour forecasting performance across multiple regions along the Texas coast and exhibits strong spatial extensibility to other coastal areas. By combining structured representation, physically grounded forcings, and scalable deep learning, this study advances the frontier of storm surge forecasting in usability, adaptability, and interpretability.