Zeyuan Hu

Sungduk Yu, Zeyuan Hu, Akshay Subramaniam et al.

Modern climate projections lack adequate spatial and temporal resolution due to computational constraints, leading to inaccuracies in representing critical processes like thunderstorms that occur on the sub-resolution scale. Hybrid methods combining physics with machine learning (ML) offer faster, higher fidelity climate simulations by outsourcing compute-hungry, high-resolution simulations to ML emulators. However, these hybrid ML-physics simulations require domain-specific data and workflows that have been inaccessible to many ML experts. As an extension of the ClimSim dataset (Yu et al., 2024), we present ClimSim-Online, which also includes an end-to-end workflow for developing hybrid ML-physics simulators. The ClimSim dataset includes 5.7 billion pairs of multivariate input/output vectors, capturing the influence of high-resolution, high-fidelity physics on a host climate simulator's macro-scale state. The dataset is global and spans ten years at a high sampling frequency. We provide a cross-platform, containerized pipeline to integrate ML models into operational climate simulators for hybrid testing. We also implement various ML baselines, alongside a hybrid baseline simulator, to highlight the ML challenges of building stable, skillful emulators. The data (https://huggingface.co/datasets/LEAP/ClimSim_high-res) and code (https://leap-stc.github.io/ClimSim and https://github.com/leap-stc/climsim-online) are publicly released to support the development of hybrid ML-physics and high-fidelity climate simulations.

Jerry Lin, Sungduk Yu, Liran Peng et al.

Machine-learning (ML) parameterizations of subgrid processes (here of turbulence, convection, and radiation) may one day replace conventional parameterizations by emulating high-resolution physics without the cost of explicit simulation. However, uncertainty about the relationship between offline and online performance (i.e., when integrated with a large-scale general circulation model (GCM)) hinders their development. Much of this uncertainty stems from limited sampling of the noisy, emergent effects of upstream ML design decisions on downstream online hybrid simulation. Our work rectifies the sampling issue via the construction of a semi-automated, end-to-end pipeline for $\mathcal{O}(100)$ size ensembles of hybrid simulations, revealing important nuances in how systematic reductions in offline error manifest in changes to online error and online stability. For example, removing dropout and switching from a Mean Squared Error (MSE) to a Mean Absolute Error (MAE) loss both reduce offline error, but they have opposite effects on online error and online stability. Other design decisions, like incorporating memory, converting moisture input from specific humidity to relative humidity, using batch normalization, and training on multiple climates do not come with any such compromises. Finally, we show that ensemble sizes of $\mathcal{O}(100)$ may be necessary to reliably detect causally relevant differences online. By enabling rapid online experimentation at scale, we can empirically settle debates regarding subgrid ML parameterization design that would have otherwise remained unresolved in the noise.

Zeyuan Hu

Equipping query processing systems with provable theoretical guarantees has been a central focus at the intersection of database theory and systems in recent years. However, the divergence between theoretical abstractions and system assumptions creates a gap between an algorithm's high-level logical specification and its low-level physical implementation. Ensuring the correctness of this logical-to-physical translation is crucial for realizing theoretical optimality as practical performance gains. Existing database testing frameworks struggle to address this need because necessary algorithm-specific inputs such as join trees are absent from standard test case generation, and integrating complex algorithms into these frameworks imposes prohibitive engineering overhead. Fallback solutions, such as using macro-benchmark queries, are inherently too noisy for isolating intricate defects during this translation. In this experience paper, we present a retrospective analysis of $\mathsf{CODA}$, a computer-orchestrated testing framework utilized during the physical co-design of TreeTracker Join ($\mathsf{TTJ}$), a theoretically optimal yet practical join algorithm recently published in ACM TODS. By synthesizing minimal reproducible examples, $\mathsf{CODA}$ successfully isolates subtle translation defects, such as state mismanagement and mapping conflicts between join trees and bushy plans. We demonstrate that this logical-to-physical translation process is a bidirectional feedback loop: early structural testing not only hardened $\mathsf{TTJ}$'s physical implementation but also exposed a boundary condition that directly refined the formal precondition of $\mathsf{TTJ}$ itself. Finally, we detail how confronting these translation challenges drove the architectural evolution of $\mathsf{CODA}$ into a robust, structure-aware test generation pipeline for join-tree-dependent algorithms.

Jinsong Yang, Zeyuan Hu, Yichen Li et al.

Underwater fish detection (UFD) is a core capability for smart aquaculture and marine ecological monitoring. While recent detectors improve accuracy by stacking feature extractors or introducing heavy attention modules, they often incur substantial computational overhead and, more importantly, neglect the physics that fundamentally limits UFD: wavelength-dependent absorption and turbidity-induced scattering significantly degrade contrast, blur fine structures, and introduce backscattering noise, leading to unreliable localization and recognition. To address these challenges, we propose FinSight-Net, an efficient and physics-aware detection framework tailored for complex aquaculture environments. FinSight-Net introduces a Multi-Scale Decoupled Dual-Stream Processing (MS-DDSP) bottleneck that explicitly targets frequency-specific information loss via heterogeneous convolutional branches, suppressing backscattering artifacts while compensating distorted biological cues through scale-aware and channel-weighted pathways. We further design an Efficient Path Aggregation FPN (EPA-FPN) as a detail-filling mechanism: it restores high-frequency spatial information typically attenuated in deep layers by establishing long-range skip connections and pruning redundant fusion routes, enabling robust detection of non-rigid fish targets under severe blur and turbidity. Extensive experiments on DeepFish, AquaFishSet, and our challenging UW-BlurredFish benchmark demonstrate that FinSight-Net achieves state-of-the-art performance. In particular, on UW-BlurredFish, FinSight-Net reaches 92.8% mAP, outperforming YOLOv11s by 4.8% while reducing parameters by 29.0%, providing a strong and lightweight solution for real-time automated monitoring in smart aquaculture.

Jerry Lin, Zeyuan Hu, Tom Beucler et al.

Subgrid machine-learning (ML) parameterizations have the potential to introduce a new generation of climate models that incorporate the effects of higher-resolution physics without incurring the prohibitive computational cost associated with more explicit physics-based simulations. However, important issues, ranging from online instability to inconsistent online performance, have limited their operational use for long-term climate projections. To more rapidly drive progress in solving these issues, domain scientists and machine learning researchers opened up the offline aspect of this problem to the broader machine learning and data science community with the release of ClimSim, a NeurIPS Datasets and Benchmarks publication, and an associated Kaggle competition. This paper reports on the downstream results of the Kaggle competition by coupling emulators inspired by the winning teams' architectures to an interactive climate model (including full cloud microphysics, a regime historically prone to online instability) and systematically evaluating their online performance. Our results demonstrate that online stability in the low-resolution, real-geography setting is reproducible across multiple diverse architectures, which we consider a key milestone. All tested architectures exhibit strikingly similar offline and online biases, though their responses to architecture-agnostic design choices (e.g., expanding the list of input variables) can differ significantly. Multiple Kaggle-inspired architectures achieve state-of-the-art (SOTA) results on certain metrics such as zonal mean bias patterns and global RMSE, indicating that crowdsourcing the essence of the offline problem is one path to improving online performance in hybrid physics-AI climate simulation.

Jinsong Yang, Zeyuan Hu, Yichen Li

Underwater fish detection (UFD) remains a challenging task in computer vision due to low object resolution, significant background interference, and high visual similarity between targets and surroundings. Existing approaches primarily focus on local feature enhancement or incorporate complex attention mechanisms to highlight small objects, often at the cost of increased model complexity and reduced efficiency. To address these limitations, we propose an efficient path aggregation network (EPANet), which leverages complementary feature integration to achieve accurate and lightweight UFD. EPANet consists of two key components: an efficient path aggregation feature pyramid network (EPA-FPN) and a multi-scale diverse-division short path bottleneck (MS-DDSP bottleneck). The EPA-FPN introduces long-range skip connections across disparate scales to improve semantic-spatial complementarity, while cross-layer fusion paths are adopted to enhance feature integration efficiency. The MS-DDSP bottleneck extends the conventional bottleneck structure by introducing finer-grained feature division and diverse convolutional operations, thereby increasing local feature diversity and representation capacity. Extensive experiments on benchmark UFD datasets demonstrate that EPANet outperforms state-of-the-art methods in terms of detection accuracy and inference speed, while maintaining comparable or even lower parameter complexity.

Zeyuan Hu, Ziang Yuan

The detection of malicious websites has become a critical issue in cybersecurity. Therefore, this paper offers a comprehensive review of data-driven methods for detecting malicious websites. Traditional approaches and their limitations are discussed, followed by an overview of data-driven approaches. The paper establishes the data-feature-model-extension pipeline and the latest research developments of data-driven approaches, including data preprocessing, feature extraction, model construction and technology extension. Specifically, this paper compares methods using deep learning models proposed in recent years. Furthermore, the paper follows the data-feature-model-extension pipeline to discuss the challenges together with some future directions of data-driven methods in malicious website detection.

Jialin Wu, Zeyuan Hu, Raymond J. Mooney

Visual question answering (VQA) and image captioning require a shared body of general knowledge connecting language and vision. We present a novel approach to improve VQA performance that exploits this connection by jointly generating captions that are targeted to help answer a specific visual question. The model is trained using an existing caption dataset by automatically determining question-relevant captions using an online gradient-based method. Experimental results on the VQA v2 challenge demonstrates that our approach obtains state-of-the-art VQA performance (e.g. 68.4% on the Test-standard set using a single model) by simultaneously generating question-relevant captions.

Jialin Wu, Zeyuan Hu, Raymond J. Mooney

Answering visual questions need acquire daily common knowledge and model the semantic connection among different parts in images, which is too difficult for VQA systems to learn from images with the only supervision from answers. Meanwhile, image captioning systems with beam search strategy tend to generate similar captions and fail to diversely describe images. To address the aforementioned issues, we present a system to have these two tasks compensate with each other, which is capable of jointly producing image captions and answering visual questions. In particular, we utilize question and image features to generate question-related captions and use the generated captions as additional features to provide new knowledge to the VQA system. For image captioning, our system attains more informative results in term of the relative improvements on VQA tasks as well as competitive results using automated metrics. Applying our system to the VQA tasks, our results on VQA v2 dataset achieve 65.8% using generated captions and 69.1% using annotated captions in validation set and 68.4% in the test-standard set. Further, an ensemble of 10 models results in 69.7% in the test-standard split.

Zeyuan Hu, Julia Strout

The goal of our research is to contribute information about how useful the crowd is at anticipating stereotypes that may be biasing a data set without a researcher's knowledge. The results of the crowd's prediction can potentially be used during data collection to help prevent the suspected stereotypes from introducing bias to the dataset. We conduct our research by asking the crowd on Amazon's Mechanical Turk (AMT) to complete two similar Human Intelligence Tasks (HITs) by suggesting stereotypes relating to their personal experience. Our analysis of these responses focuses on determining the level of diversity in the workers' suggestions and their demographics. Through this process we begin a discussion on how useful the crowd can be in tackling this difficult problem within machine learning data collection.