arXiv subjects

Gang Hu

Publications and source records attributed to Gang Hu.

At least 19 recordsLinked to original sources

arXiv2026

TaxPraBen: A Scalable Benchmark for Structured Evaluation of LLMs in Chinese Real-World Tax Practice

While Large Language Models (LLMs) excel in various general domains, they exhibit notable gaps in the highly specialized, knowledge-intensive, and legally regulated Chinese tax domain. Consequently, while tax-related benchmarks are gaining attention, many focus on isolated NLP tasks, neglecting real-world practical capabilities. To address this issue, we introduce TaxPraBen, the first dedicated benchmark for Chinese taxation practice. It combines 10 traditional application tasks, along with 3 pioneering real-world scenarios: tax risk prevention, tax inspection analysis, and tax strategy planning, sourced from 14 datasets totaling 7.3K instances. TaxPraBen features a scalable structured evaluation paradigm designed through process of "structured parsing-field alignment extraction-numerical and textual matching", enabling end-to-end tax practice assessment while being extensible to other domains. We evaluate 19 LLMs based on Bloom's taxonomy. The results indicate significant performance disparities: all closed-source large-parameter LLMs excel, and Chinese LLMs like Qwen2.5 generally exceed multilingual LLMs, while the YaYi2 LLM, fine-tuned with some tax data, shows only limited improvement. TaxPraBen serves as a vital resource for advancing evaluations of LLMs in practical applications.

cs.CL↗

arXiv2026

XBRLTagRec: Domain-Specific Fine-Tuning and Zero-Shot Re-Ranking with LLMs for Extreme Financial Numeral Labeling

Publicly traded companies must disclose financial information under regulations of the Securities and Exchange Commission (SEC) and the Generally Accepted Accounting Principles (GAAP). The eXtensible Business Reporting Language (XBRL), as an XML-based financial language, enables standardized and machine-readable reporting, but accurate tag selection from large taxonomies remains challenging. Existing fine-tuning-based methods struggle to distinguish highly similar XBRL tags, limiting performance in financial data matching. To address these issues, we introduce XBRLTagRec, an end-to-end framework for automated financial numeral tagging. The framework generates semantic tag documents with a fine-tuned FLAN-T5-Large model, retrieves relevant candidates via semantic similarity, and applies zero-shot re-ranking with ChatGPT-3.5 to select the optimal tag. Experiments on the FNXL dataset show that XBRLTagRec outperforms the state-of-the-art FLAN-FinXC framework, achieving 2.64%-4.47% improvements in Hits@1 and Macro metrics. These results demonstrate its effectiveness in large-scale and semantically complex tag matching scenarios.

cs.CE↗

arXiv2026

CA-HFP: Curvature-Aware Heterogeneous Federated Pruning with Model Reconstruction

Federated learning on heterogeneous edge devices requires personalized compression while preserving aggregation compatibility and stable convergence. We present Curvature-Aware Heterogeneous Federated Pruning (CA-HFP), a practical framework that enables each client perform structured, device-specific pruning guided by a curvature-informed significance score, and subsequently maps its compact submodel back into a common global parameter space via a lightweight reconstruction. We derive a convergence bound for federated optimization with multiple local SGD steps that explicitly accounts for local computation, data heterogeneity, and pruning-induced perturbations; from which a principled loss-based pruning criterion is derived. Extensive experiments on FMNIST, CIFAR-10, and CIFAR-100 using VGG and ResNet architectures under varying degrees of data heterogeneity demonstrate that CA-HFP preserves model accuracy while significantly reducing per-client computation and communication costs, outperforming standard federated training and existing pruning-based baselines.

cs.LG↗

arXiv2025

Optimization of lenslet arrays for PRIMA Kinetic Inductance Detectors

The PRobe far-Infrared Mission for Astrophysics (PRIMA) is a cryogenically cooled 1.8-m space telescope designed to address fundamental questions about the evolution of galactic ecosystems, the origins of planetary atmospheres, and the buildup of dust and metals over cosmic time. PRIMA will achieve unprecedented sensitivity in the 24 - 261 $\mu$m wavelength range, enabled by background-limited kinetic inductance detectors (KIDs) cooled to 120 mK. For PRIMA's Far-InfraRed Enhanced Survey Spectrometer (FIRESS) instrument, we have developed monolithic kilopixel silicon lenslet arrays to efficiently couple incident radiation from the telescope's fore-optics onto the KID absorber elements. These three-dimensional lenslet arrays are fabricated using grayscale lithography, followed by deep reactive ion etching (DRIE), and are anti-reflection (AR) coated with a quarter-wavelength thick deposition of Parylene-C. The lenslet arrays are aligned and bonded to the KID arrays using a thin layer of epoxy through a flip-chip bonder. In this work, we report on the optimized fabrication, lens design, AR coating, and bonding processes developed for the FIRESS lenslet arrays. We characterize brassboard lenslet arrays fabricated to meet the specifications of the FIRESS low and high spectral bands, demonstrate stepped-thickness AR-coatings to achieve high efficiency across broad wavelength ranges, and present spectral transmission measurements of the AR coating and the epoxy bonding layers.

astro-ph.IM↗

arXiv2025

FFT-MoE: Efficient Federated Fine-Tuning for Foundation Models via Large-scale Sparse MoE under Heterogeneous Edge

As FMs drive progress toward Artificial General Intelligence (AGI), fine-tuning them under privacy and resource constraints has become increasingly critical particularly when highquality training data resides on distributed edge devices. Federated Learning (FL) offers a compelling solution through Federated Fine-Tuning (FFT), which enables collaborative model adaptation without sharing raw data. Recent approaches incorporate Parameter-Efficient Fine-Tuning (PEFT) techniques such as Low Rank Adaptation (LoRA) to reduce computational overhead. However, LoRA-based FFT faces two major limitations in heterogeneous FL environments: structural incompatibility across clients with varying LoRA configurations and limited adaptability to non-IID data distributions, which hinders convergence and generalization. To address these challenges, we propose FFT MoE, a novel FFT framework that replaces LoRA with sparse Mixture of Experts (MoE) adapters. Each client trains a lightweight gating network to selectively activate a personalized subset of experts, enabling fine-grained adaptation to local resource budgets while preserving aggregation compatibility. To further combat the expert load imbalance caused by device and data heterogeneity, we introduce a heterogeneity-aware auxiliary loss that dynamically regularizes the routing distribution to ensure expert diversity and balanced utilization. Extensive experiments spanning both IID and non-IID conditions demonstrate that FFT MoE consistently outperforms state of the art FFT baselines in generalization performance and training efficiency.

cs.LG↗

arXiv2025

LatentFlow: Cross-Frequency Experimental Flow Reconstruction from Sparse Pressure via Latent Mapping

Acquiring temporally high-frequency and spatially high-resolution turbulent wake flow fields in particle image velocimetry (PIV) experiments remains a significant challenge due to hardware limitations and measurement noise. In contrast, temporal high-frequency measurements of spatially sparse wall pressure are more readily accessible in wind tunnel experiments. In this study, we propose a novel cross-modal temporal upscaling framework, LatentFlow, which reconstructs high-frequency (512 Hz) turbulent wake flow fields by fusing synchronized low-frequency (15 Hz) flow field and pressure data during training, and high-frequency wall pressure signals during inference. The first stage involves training a pressure-conditioned $\beta$-variation autoencoder ($p$C-$\beta$-VAE) to learn a compact latent representation that captures the intrinsic dynamics of the wake flow. A secondary network maps synchronized low-frequency wall pressure signals into the latent space, enabling reconstruction of the wake flow field solely from sparse wall pressure. Once trained, the model utilizes high-frequency, spatially sparse wall pressure inputs to generate corresponding high-frequency flow fields via the $p$C-$\beta$-VAE decoder. By decoupling the spatial encoding of flow dynamics from temporal pressure measurements, LatentFlow provides a scalable and robust solution for reconstructing high-frequency turbulent wake flows in data-constrained experimental settings.

cs.LG↗

arXiv2025

Spatiotemporal wall pressure forecast of a rectangular cylinder with physics-aware DeepU-Fourier neural network

The wall pressure is of great importance in understanding the forces and structural responses induced by fluid. Recent works have investigated the potential of deep learning techniques in predicting mean pressure coefficients and fluctuating pressure coefficients, but most of existing deep learning frameworks are limited to predicting a single snapshot using full spatial information. To forecast spatiotemporal wall pressure of flow past a rectangular cylinder, this study develops a physics-aware DeepU-Fourier neural Network (DeepUFNet) deep learning model. DeepUFNet comprises the UNet structure and the Fourier neural network, with physical high-frequency loss control embedded in the model training stage to optimize model performance. Wind tunnel testing was performed to collect wall pressures on two-dimensional rectangular cylinders using high-frequency pressure scanning, thereby constructing a database for DeepUFNet training and testing. The DeepUFNet model is found capable of forecasting spatiotemporal wall pressure information with high accuracy on the rectangular cylinder with side ratio 1.5. The comparison between forecast results and experimental data presents agreement in statistical information and physical interpretation. It is also found that embedding a physical high-frequency loss control coefficient b in the DeepUFNet model can significantly improve model performance in forecasting spatiotemporal wall pressure information, particularly, high-order frequency fluctuation and wall pressure variance. Furthermore, the DeepUFNet extrapolation capability is tested with sparse spatial information input, and the model presents a satisfactory extrapolation ability. Last, the DeepUFNet is tested for generalization in unseen cases, rectangular cylinders with side ratio 4 and 3.75, and the model presents satisfactory generalization ability.

physics.flu-dyn↗

arXiv2024

Split Federated Learning Over Heterogeneous Edge Devices: Algorithm and Optimization

Split Learning (SL) is a promising collaborative machine learning approach, enabling resource-constrained devices to train models without sharing raw data, while reducing computational load and preserving privacy simultaneously. However, current SL algorithms face limitations in training efficiency and suffer from prolonged latency, particularly in sequential settings, where the slowest device can bottleneck the entire process due to heterogeneous resources and frequent data exchanges between clients and servers. To address these challenges, we propose the Heterogeneous Split Federated Learning (HSFL) framework, which allows resource-constrained clients to train their personalized client-side models in parallel, utilizing different cut layers. Aiming to mitigate the impact of heterogeneous environments and accelerate the training process, we formulate a latency minimization problem that optimizes computational and transmission resources jointly. Additionally, we design a resource allocation algorithm that combines the Sample Average Approximation (SAA), Genetic Algorithm (GA), Lagrangian relaxation and Branch and Bound (B\&B) methods to efficiently solve this problem. Simulation results demonstrate that HSFL outperforms other frameworks in terms of both convergence rate and model accuracy on heterogeneous devices with non-iid data, while the optimization algorithm is better than other baseline methods in reducing latency.

cs.LG↗

arXiv2024

Machine-learned flow estimation with sparse data -- exemplified for the rooftop of a UAV vertiport

We propose a physics-informed data-driven framework for urban wind estimation. This framework validates and incorporates the Reynolds number independence for flows under various working conditions, thus allowing the extrapolation for wind conditions far beyond the training data. Another key enabler is a machine-learned non-dimensionalized manifold from snapshot data. The velocity field is modeled using a double encoder-decoder approach. The first encoder normalizes data using the oncoming wind speed, while the second encoder projects this normalized data onto the isometric feature mapping manifold. The decoders reverse this process, with $k$-nearest neighbor performing the first decoding and the second undoing the normalization. The manifold is coarse-grained by clustering to reduce the computational load for de- and encoding. The sensor-based flow estimation is based on the estimate of the oncoming wind speed and a mapping from sensor signal to the manifold latent variables. The proposed machine-learned flow estimation framework is exemplified for the flow above an Unmanned Aerial Vehicle vertiport. The wind estimation is shown to generalize well for rare wind conditions, not included in the original database.

physics.flu-dyn↗

arXiv2024

Mutagenesis screen to map the functions of parameters of Large Language Models

Large Language Models (LLMs) have significantly advanced artificial intelligence, excelling in numerous tasks. Although the functionality of a model is inherently tied to its parameters, a systematic method for exploring the connections between the parameters and the functionality are lacking. Models sharing similar structure and parameter counts exhibit significant performance disparities across various tasks, prompting investigations into the varying patterns that govern their performance. We adopted a mutagenesis screen approach inspired by the methods used in biological studies, to investigate Llama2-7b and Zephyr. This technique involved mutating elements within the models' matrices to their maximum or minimum values to examine the relationship between model parameters and their functionalities. Our research uncovered multiple levels of fine structures within both models. Many matrices showed a mixture of maximum and minimum mutations following mutagenesis, but others were predominantly sensitive to one type. Notably, mutations that produced phenotypes, especially those with severe outcomes, tended to cluster along axes. Additionally, the location of maximum and minimum mutations often displayed a complementary pattern on matrix in both models, with the Gate matrix showing a unique two-dimensional asymmetry after rearrangement. In Zephyr, certain mutations consistently resulted in poetic or conversational rather than descriptive outputs. These "writer" mutations grouped according to the high-frequency initial word of the output, with a marked tendency to share the row coordinate even when they are in different matrices. Our findings affirm that the mutagenesis screen is an effective tool for deciphering the complexities of large language models and identifying unexpected ways to expand their potential, providing deeper insights into the foundational aspects of AI systems.

cs.AI↗

arXiv2024

Faster Convergence on Heterogeneous Federated Edge Learning: An Adaptive Clustered Data Sharing Approach

Federated Edge Learning (FEEL) emerges as a pioneering distributed machine learning paradigm for the 6G Hyper-Connectivity, harnessing data from the Internet of Things (IoT) devices while upholding data privacy. However, current FEEL algorithms struggle with non-independent and non-identically distributed (non-IID) data, leading to elevated communication costs and compromised model accuracy. To address these statistical imbalances within FEEL, we introduce a clustered data sharing framework, mitigating data heterogeneity by selectively sharing partial data from cluster heads to trusted associates through sidelink-aided multicasting. The collective communication pattern is integral to FEEL training, where both cluster formation and the efficiency of communication and computation impact training latency and accuracy simultaneously. To tackle the strictly coupled data sharing and resource optimization, we decompose the overall optimization problem into the clients clustering and effective data sharing subproblems. Specifically, a distribution-based adaptive clustering algorithm (DACA) is devised basing on three deductive cluster forming conditions, which ensures the maximum sharing yield. Meanwhile, we design a stochastic optimization based joint computed frequency and shared data volume optimization (JFVO) algorithm, determining the optimal resource allocation with an uncertain objective function. The experiments show that the proposed framework facilitates FEEL on non-IID datasets with faster convergence rate and higher model accuracy in a limited communication environment.

cs.LG↗

arXiv2024

Markowitz Meets Bellman: Knowledge-distilled Reinforcement Learning for Portfolio Management

Investment portfolios, central to finance, balance potential returns and risks. This paper introduces a hybrid approach combining Markowitz's portfolio theory with reinforcement learning, utilizing knowledge distillation for training agents. In particular, our proposed method, called KDD (Knowledge Distillation DDPG), consist of two training stages: supervised and reinforcement learning stages. The trained agents optimize portfolio assembly. A comparative analysis against standard financial models and AI frameworks, using metrics like returns, the Sharpe ratio, and nine evaluation indices, reveals our model's superiority. It notably achieves the highest yield and Sharpe ratio of 2.03, ensuring top profitability with the lowest risk in comparable return scenarios.

q-fin.CP↗

arXiv2024

No Language is an Island: Unifying Chinese and English in Financial Large Language Models, Instruction Data, and Benchmarks

While the progression of Large Language Models (LLMs) has notably propelled financial analysis, their application has largely been confined to singular language realms, leaving untapped the potential of bilingual Chinese-English capacity. To bridge this chasm, we introduce ICE-PIXIU, seamlessly amalgamating the ICE-INTENT model and ICE-FLARE benchmark for bilingual financial analysis. ICE-PIXIU uniquely integrates a spectrum of Chinese tasks, alongside translated and original English datasets, enriching the breadth and depth of bilingual financial modeling. It provides unrestricted access to diverse model variants, a substantial compilation of diverse cross-lingual and multi-modal instruction data, and an evaluation benchmark with expert annotations, comprising 10 NLP tasks, 20 bilingual specific tasks, totaling 95k datasets. Our thorough evaluation emphasizes the advantages of incorporating these bilingual datasets, especially in translation tasks and utilizing original English data, enhancing both linguistic flexibility and analytical acuity in financial contexts. Notably, ICE-INTENT distinguishes itself by showcasing significant enhancements over conventional LLMs and existing financial LLMs in bilingual milieus, underscoring the profound impact of robust bilingual data on the accuracy and efficacy of financial NLP.

cs.CE↗

arXiv2024

FinBen: A Holistic Financial Benchmark for Large Language Models

LLMs have transformed NLP and shown promise in various fields, yet their potential in finance is underexplored due to a lack of comprehensive evaluation benchmarks, the rapid development of LLMs, and the complexity of financial tasks. In this paper, we introduce FinBen, the first extensive open-source evaluation benchmark, including 36 datasets spanning 24 financial tasks, covering seven critical aspects: information extraction (IE), textual analysis, question answering (QA), text generation, risk management, forecasting, and decision-making. FinBen offers several key innovations: a broader range of tasks and datasets, the first evaluation of stock trading, novel agent and Retrieval-Augmented Generation (RAG) evaluation, and three novel open-source evaluation datasets for text summarization, question answering, and stock trading. Our evaluation of 15 representative LLMs, including GPT-4, ChatGPT, and the latest Gemini, reveals several key findings: While LLMs excel in IE and textual analysis, they struggle with advanced reasoning and complex tasks like text generation and forecasting. GPT-4 excels in IE and stock trading, while Gemini is better at text generation and forecasting. Instruction-tuned LLMs improve textual analysis but offer limited benefits for complex tasks such as QA. FinBen has been used to host the first financial LLMs shared task at the FinNLP-AgentScen workshop during IJCAI-2024, attracting 12 teams. Their novel solutions outperformed GPT-4, showcasing FinBen's potential to drive innovation in financial LLMs. All datasets, results, and codes are released for the research community: https://github.com/The-FinAI/PIXIU.

cs.CL↗

arXiv2024

Deep reinforcement transfer learning for active flow control of a 3D square cylinder under state dimension mismatch

This paper focuses on developing a deep reinforcement learning (DRL) control strategy to mitigate aerodynamic forces acting on a three dimensional (3D) square cylinder under high Reynolds number flow conditions. Four jets situated at the corners of the square cylinder are used as actuators and pressure probes on the cylinder surface are employed as feedback observers. The Soft Actor-Critic (SAC) algorithm is deployed to identify an effective control scheme. Additionally, we pre-train the DRL agent using a two dimensional (2D) square cylinder flow field at a low Reynolds number (Re =1000), followed by transferring it to the 3D square cylinder at Re =22000. To address the issue of state dimension mismatch in transfer learning from 2D to 3D case, a state dimension mismatch transfer learning method is developed to enhance the SAC algorithm, named SDTL-SAC. The results demonstrate transfer learning across different state spaces achieves the same control policy as the SAC algorithm, resulting in a significant improvement in training speed with a training cost reduction of 51.1%. Furthermore, the SAC control strategy leads to a notable 52.3% reduction in drag coefficient, accompanied by substantial suppression of lift fluctuations. These outcomes underscore the potential of DRL in active flow control, laying the groundwork for efficient, robust, and practical implementation of this control technique in practical engineering.

physics.flu-dyn↗

arXiv2023

Advancing Algorithmic Trading: A Multi-Technique Enhancement of Deep Q-Network Models

This study enhances a Deep Q-Network (DQN) trading model by incorporating advanced techniques like Prioritized Experience Replay, Regularized Q-Learning, Noisy Networks, Dueling, and Double DQN. Extensive tests on assets like BTC/USD and AAPL demonstrate superior performance compared to the original model, with marked increases in returns and Sharpe Ratio, indicating improved risk-adjusted rewards. Notably, convolutional neural network (CNN) architectures, both 1D and 2D, significantly boost returns, suggesting their effectiveness in market trend analysis. Across instruments, these enhancements have yielded stable and high gains, eclipsing the baseline and highlighting the potential of CNNs in trading systems. The study suggests that applying sophisticated deep learning within reinforcement learning can greatly enhance automated trading, urging further exploration into advanced methods for broader financial applicability. The findings advocate for the continued evolution of AI in finance.

q-fin.CP↗

arXiv2023

Aerodynamic Characterization of a Fan Array Wind Generator

Experimental assessment of safe and precise flight control algorithms for unmanned aerial vehicles (UAVs) under gusty wind conditions requires the capability to generate a large range of velocity profiles. In this study, we employ a small fan array wind generator which can generate flows with large spatial and temporal variability. We perform a thorough aerodynamic characterization operating the fans uniformly from a low to the maximum level. PIV and hot-wire measurements indicate a jet-like flow with nearly uniform core which monotonously contracts in streamwise direction and surrounding growing unsteady shear-layers. These complex dynamics results in a limited region with desired flow profile and turbulence level. The experimental results shed light on the flow generated by a full-scale fan array wind generator, and indicate the need for further improvements via properly designed add-ons and dedicated control algorithms.

physics.flu-dyn↗

arXiv2023

Dynamic Feature-based Deep Reinforcement Learning for Flow Control of Circular Cylinder with Sparse Surface Pressure Sensing

This study proposes a self-learning algorithm for closed-loop cylinder wake control targeting lower drag and lower lift fluctuations with the additional challenge of sparse sensor information, taking deep reinforcement learning as the starting point. DRL performance is significantly improved by lifting the sensor signals to dynamic features (DF), which predict future flow states. The resulting dynamic feature-based DRL (DF-DRL) automatically learns a feedback control in the plant without a dynamic model. Results show that the drag coefficient of the DF-DRL model is 25% less than the vanilla model based on direct sensor feedback. More importantly, using only one surface pressure sensor, DF-DRL can reduce the drag coefficient to a state-of-the-art performance of about 8% at Re = 100 and significantly mitigate lift coefficient fluctuations. Hence, DF-DRL allows the deployment of sparse sensing of the flow without degrading the control performance. This method also shows good robustness in controlling flow under higher Reynolds numbers, which reduces the drag coefficient by 32.2% and 46.55% at Re = 500 and 1000, respectively, indicating the broad applicability of the method. Since surface pressure information is more straightforward to measure in realistic scenarios than flow velocity information, this study provides a valuable reference for experimentally designing the active flow control of a circular cylinder based on wall pressure signals, which is an essential step toward further developing intelligent control in realistic multi-input multi-output (MIMO) system.

cs.LG↗