arXiv subjects

Feifei Wang

Publications and source records attributed to Feifei Wang.

At least 37 records · Page 2Linked to original sources

arXiv2025

PrAda-GAN: A Private Adaptive Generative Adversarial Network with Bayes Network Structure

We revisit the problem of generating synthetic data under differential privacy. To address the core limitations of marginal-based methods, we propose the Private Adaptive Generative Adversarial Network with Bayes Network Structure (PrAda-GAN), which integrates the strengths of both GAN-based and marginal-based approaches. Our method adopts a sequential generator architecture to capture complex dependencies among variables, while adaptively regularizing the learned structure to promote sparsity in the underlying Bayes network. Theoretically, we establish diminishing bounds on the parameter distance, variable selection error, and Wasserstein distance. Our analysis shows that leveraging dependency sparsity leads to significant improvements in convergence rates. Empirically, experiments on both synthetic and real-world datasets demonstrate that PrAda-GAN outperforms existing tabular data synthesis methods in terms of the privacy-utility trade-off.

stat.ML↗

arXiv2025

Doubly Smoothed Density Estimation with Application on Miners' Unsafe Act Detection

We study anomaly detection in images under a fixed-camera environment and propose a \emph{doubly smoothed} (DS) density estimator that exploits spatial structure to improve estimation accuracy. The DS estimator applies kernel smoothing twice: first over the value domain to obtain location-wise classical nonparametric density (CD) estimates, and then over the spatial domain to borrow information from neighboring locations. Under appropriate regularity conditions, we show that the DS estimator achieves smaller asymptotic bias, variance, and mean squared error than the CD estimator. To address the increased computational cost of the DS estimator, we introduce a grid point approximation (GPA) technique that reduces the computation cost of inference without sacrificing the estimation accuracy. A rule-of-thumb bandwidth is derived for practical use. Extensive simulations show that GPA-DS achieves the lowest MSE with near real-time speed. In a large-scale case study on underground mine surveillance, GPA-DS enables remarkable sub-image extraction of anomalous regions after which a lightweight MobileNet classifier achieves $\approx$99\% out-of-sample accuracy for unsafe act detection.

stat.AP↗

arXiv2025

A BERT-based Hierarchical Classification Model with Applications in Chinese Commodity Classification

Existing e-commerce platforms heavily rely on manual annotation for product categorization, which is inefficient and inconsistent. These platforms often employ a hierarchical structure for categorizing products; however, few studies have leveraged this hierarchical information for classification. Furthermore, studies that consider hierarchical information fail to account for similarities and differences across various hierarchical categories. Herein, we introduce a large-scale hierarchical dataset collected from the JD e-commerce platform (www.JD.com), comprising 1,011,450 products with titles and a three-level category structure. By making this dataset openly accessible, we provide a valuable resource for researchers and practitioners to advance research and applications associated with product categorization. Moreover, we propose a novel hierarchical text classification approach based on the widely used Bidirectional Encoder Representations from Transformers (BERT), called Hierarchical Fine-tuning BERT (HFT-BERT). HFT-BERT leverages the remarkable text feature extraction capabilities of BERT, achieving prediction performance comparable to those of existing methods on short texts. Notably, our HFT-BERT model demonstrates exceptional performance in categorizing longer short texts, such as books.

cs.CL↗

arXiv2025

FedVLMBench: Benchmarking Federated Fine-Tuning of Vision-Language Models

Vision-Language Models (VLMs) have demonstrated remarkable capabilities in cross-modal understanding and generation by integrating visual and textual information. While instruction tuning and parameter-efficient fine-tuning methods have substantially improved the generalization of VLMs, most existing approaches rely on centralized training, posing challenges for deployment in domains with strict privacy requirements like healthcare. Recent efforts have introduced Federated Learning (FL) into VLM fine-tuning to address these privacy concerns, yet comprehensive benchmarks for evaluating federated fine-tuning strategies, model architectures, and task generalization remain lacking. In this work, we present \textbf{FedVLMBench}, the first systematic benchmark for federated fine-tuning of VLMs. FedVLMBench integrates two mainstream VLM architectures (encoder-based and encoder-free), four fine-tuning strategies, five FL algorithms, six multimodal datasets spanning four cross-domain single-task scenarios and two cross-domain multitask settings, covering four distinct downstream task categories. Through extensive experiments, we uncover key insights into the interplay between VLM architectures, fine-tuning strategies, data heterogeneity, and multi-task federated optimization. Notably, we find that a 2-layer multilayer perceptron (MLP) connector with concurrent connector and LLM tuning emerges as the optimal configuration for encoder-based VLMs in FL. Furthermore, current FL methods exhibit significantly higher sensitivity to data heterogeneity in vision-centric tasks than text-centric ones, across both encoder-free and encoder-based VLM architectures. Our benchmark provides essential tools, datasets, and empirical guidance for the research community, offering a standardized platform to advance privacy-preserving, federated training of multimodal foundation models.

cs.LG↗

arXiv2025

Selective Aggregation for Low-Rank Adaptation in Federated Learning

We investigate LoRA in federated learning through the lens of the asymmetry analysis of the learned $A$ and $B$ matrices. In doing so, we uncover that $A$ matrices are responsible for learning general knowledge, while $B$ matrices focus on capturing client-specific knowledge. Based on this finding, we introduce Federated Share-A Low-Rank Adaptation (FedSA-LoRA), which employs two low-rank trainable matrices $A$ and $B$ to model the weight update, but only $A$ matrices are shared with the server for aggregation. Moreover, we delve into the relationship between the learned $A$ and $B$ matrices in other LoRA variants, such as rsLoRA and VeRA, revealing a consistent pattern. Consequently, we extend our FedSA-LoRA method to these LoRA variants, resulting in FedSA-rsLoRA and FedSA-VeRA. In this way, we establish a general paradigm for integrating LoRA with FL, offering guidance for future work on subsequent LoRA variants combined with FL. Extensive experimental results on natural language understanding and generation tasks demonstrate the effectiveness of the proposed method. Our code is available at https://github.com/Pengxin-Guo/FedSA-LoRA.

cs.LG↗

arXiv2025

Auxiliary Learning and its Statistical Understanding

Modern statistical analysis often encounters high-dimensional problems but with a limited sample size. It poses great challenges to traditional statistical estimation methods. In this work, we adopt auxiliary learning to solve the estimation problem in high-dimensional settings. We start with the linear regression setup. To improve the statistical efficiency of the parameter estimator for the primary task, we consider several auxiliary tasks, which share the same covariates with the primary task. Then a weighted estimator for the primary task is developed, which is a linear combination of the ordinary least squares estimators of both the primary task and auxiliary tasks. The optimal weight is analytically derived and the statistical properties of the corresponding weighted estimator are studied. We then extend the weighted estimator to generalized linear regression models. Extensive numerical experiments are conducted to verify our theoretical results. Last, a deep learning-related real-data example of smart vending machines is presented for illustration purposes.

math.ST↗

arXiv2024

MambaClinix: Hierarchical Gated Convolution and Mamba-Based U-Net for Enhanced 3D Medical Image Segmentation

Deep learning, particularly convolutional neural networks (CNNs) and Transformers, has significantly advanced 3D medical image segmentation. While CNNs are highly effective at capturing local features, their limited receptive fields may hinder performance in complex clinical scenarios. In contrast, Transformers excel at modeling long-range dependencies but are computationally intensive, making them expensive to train and deploy. Recently, the Mamba architecture, based on the State Space Model (SSM), has been proposed to efficiently model long-range dependencies while maintaining linear computational complexity. However, its application in medical image segmentation reveals shortcomings, particularly in capturing critical local features essential for accurate delineation of clinical regions. In this study, we propose MambaClinix, a novel U-shaped architecture for medical image segmentation that integrates a hierarchical gated convolutional network(HGCN) with Mamba in an adaptive stage-wise framework. This design significantly enhances computational efficiency and high-order spatial interactions, enabling the model to effectively capture both proximal and distal relationships in medical images. Specifically, our HGCN is designed to mimic the attention mechanism of Transformers by a purely convolutional structure, facilitating high-order spatial interactions in feature maps while avoiding the computational complexity typically associated with Transformer-based methods. Additionally, we introduce a region-specific Tversky loss, which emphasizes specific pixel regions to improve auto-segmentation performance, thereby optimizing the model's decision-making process. Experimental results on five benchmark datasets demonstrate that the proposed MambaClinix achieves high segmentation accuracy while maintaining low model complexity.

eess.IV↗

arXiv2024

Face Swap via Diffusion Model

This technical report presents a diffusion model based framework for face swapping between two portrait images. The basic framework consists of three components, i.e., IP-Adapter, ControlNet, and Stable Diffusion's inpainting pipeline, for face feature encoding, multi-conditional generation, and face inpainting respectively. Besides, I introduce facial guidance optimization and CodeFormer based blending to further improve the generation quality. Specifically, we engage a recent light-weighted customization method (i.e., DreamBooth-LoRA), to guarantee the identity consistency by 1) using a rare identifier "sks" to represent the source identity, and 2) injecting the image features of source portrait into each cross-attention layer like the text features. Then I resort to the strong inpainting ability of Stable Diffusion, and utilize canny image and face detection annotation of the target portrait as the conditions, to guide ContorlNet's generation and align source portrait with the target portrait. To further correct face alignment, we add the facial guidance loss to optimize the text embedding during the sample generation. The code is available at: https://github.com/somuchtome/Faceswap

cs.CV↗

arXiv2024

A Selective Review on Statistical Methods for Massive Data Computation: Distributed Computing, Subsampling, and Minibatch Techniques

This paper presents a selective review of statistical computation methods for massive data analysis. A huge amount of statistical methods for massive data computation have been rapidly developed in the past decades. In this work, we focus on three categories of statistical computation methods: (1) distributed computing, (2) subsampling methods, and (3) minibatch gradient techniques. The first class of literature is about distributed computing and focuses on the situation, where the dataset size is too huge to be comfortably handled by one single computer. In this case, a distributed computation system with multiple computers has to be utilized. The second class of literature is about subsampling methods and concerns about the situation, where the sample size of dataset is small enough to be placed on one single computer but too large to be easily processed by its memory as a whole. The last class of literature studies those minibatch gradient related optimization techniques, which have been extensively used for optimizing various deep learning models.

stat.ME↗

arXiv2024

Cooperation and interdependence in global science funding

Investments in research and development are key to scientific and economic growth and to the well-being of society. Scientific research demands significant resources making national scientific investment a crucial driver of scientific production. As scientific production becomes increasingly multinational, it is critical to study how nations' scientific activities are funded both domestically and internationally. By tracing research grants acknowledged in scholarly publications, our study reveals a shifting duopoly of China and the United States in the global funding landscape, with a contrasting funding pattern; while China has surpassed the United States in publications with acknowledged domestic and international funding, the United States largely maintains its role as the most important global research partner. Our results also highlight the precarity of low- and middle-income countries to global funding disruptions. By revealing the complex interdependence and collaboration between countries in the global scientific enterprise, this work informs future studies investigating the national and global scientific enterprise and how funding leads to both productive cooperation and vulnerable dependencies.

econ.GN↗

arXiv2023

Hierarchical Terrain Attention and Multi-Scale Rainfall Guidance For Flood Image Prediction

With the deterioration of climate, the phenomenon of rain-induced flooding has become frequent. To mitigate its impact, recent works adopt convolutional neural network or its variants to predict the floods. However, these methods directly force the model to reconstruct the raw pixels of flood images through a global constraint, overlooking the underlying information contained in terrain features and rainfall patterns. To address this, we present a novel framework for precise flood map prediction, which incorporates hierarchical terrain spatial attention to help the model focus on spatially-salient areas of terrain features and constructs multi-scale rainfall embedding to extensively integrate rainfall pattern information into generation. To better adapt the model in various rainfall conditions, we leverage a rainfall regression loss for both the generator and the discriminator as additional supervision. Extensive evaluations on real catchment datasets demonstrate the superior performance of our method, which greatly surpasses the previous arts under different rainfall conditions.

cs.CV↗

arXiv2023

Deep Learning Enables Large Depth-of-Field Images for Sub-Diffraction-Limit Scanning Superlens Microscopy

Scanning electron microscopy (SEM) is indispensable in diverse applications ranging from microelectronics to food processing because it provides large depth-of-field images with a resolution beyond the optical diffraction limit. However, the technology requires coating conductive films on insulator samples and a vacuum environment. We use deep learning to obtain the mapping relationship between optical super-resolution (OSR) images and SEM domain images, which enables the transformation of OSR images into SEM-like large depth-of-field images. Our custom-built scanning superlens microscopy (SSUM) system, which requires neither coating samples by conductive films nor a vacuum environment, is used to acquire the OSR images with features down to ~80 nm. The peak signal-to-noise ratio (PSNR) and structural similarity index measure values indicate that the deep learning method performs excellently in image-to-image translation, with a PSNR improvement of about 0.74 dB over the optical super-resolution images. The proposed method provides a high level of detail in the reconstructed results, indicating that it has broad applicability to chip-level defect detection, biological sample analysis, forensics, and various other fields.

physics.optics↗

arXiv2023

Rechargeable Li/Cl$_2$ battery down to -80 °C

Low temperature rechargeable batteries are important to life in cold climates, polar/deep-sea expeditions and space explorations. Here, we report ~ 3.5 - 4 V rechargeable lithium/chlorine (Li/Cl2) batteries operating down to -80 °C, employing Li metal negative electrode, a novel CO2 activated porous carbon (KJCO2) as the positive electrode, and a high ionic conductivity (~ 5 to 20 mS cm-1 from -80 °C to 25 °C) electrolyte comprised of 1 M aluminum chloride (AlCl3), 0.95 M lithium chloride (LiCl), and 0.05 M lithium bis(fluorosulfonyl)imide (LiFSI) in low melting point (-104.5 °C) thionyl chloride (SOCl2). Between room-temperature and -80 °C, the Li/Cl2 battery delivered up to ~ 30,000 - 4,500 mAh g-1 first discharge capacity and a 1,200 - 5,000 mAh g-1 reversible capacity (discharge voltages in ~ 3.5 to 3.1 V) over up to 130 charge-discharge cycles. Mass spectrometry and X-ray photoelectron spectroscopy (XPS) probed Cl2 trapped in the porous carbon upon LiCl electro-oxidation during charging. At lower temperature down to -80 °C, SCl2/S2Cl2 and Cl2 generated by electro-oxidation in the charging step were trapped in porous KJCO2 carbon, allowing for reversible reduction to afford a high discharge voltage plateau near ~ 4 V with up to ~ 1000 mAh g-1 capacity for SCl2/S2Cl2 reduction and up to ~ 4000 mAh g-1 capacity at ~ 3.1 V plateau for Cl2 reduction. Towards practical use, we made CR2032 Li/Cl2 battery cells to drive digital watches at -40 °C and light emitting diode at -80 °C, opening Li/Cl2 secondary batteries for ultra-cold conditions.

physics.app-ph↗

arXiv2023

Subnetwork Estimation for Spatial Autoregressive Models in Large-scale Networks

Large-scale networks are commonly encountered in practice (e.g., Facebook and Twitter) by researchers. In order to study the network interaction between different nodes of large-scale networks, the spatial autoregressive (SAR) model has been popularly employed. Despite its popularity, the estimation of a SAR model on large-scale networks remains very challenging. On the one hand, due to policy limitations or high collection costs, it is often impossible for independent researchers to observe or collect all network information. On the other hand, even if the entire network is accessible, estimating the SAR model using the quasi-maximum likelihood estimator (QMLE) could be computationally infeasible due to its high computational cost. To address these challenges, we propose here a subnetwork estimation method based on QMLE for the SAR model. By using appropriate sampling methods, a subnetwork, consisting of a much-reduced number of nodes, can be constructed. Subsequently, the standard QMLE can be computed by treating the sampled subnetwork as if it were the entire network. This leads to a significant reduction in information collection and model computation costs, which increases the practical feasibility of the effort. Theoretically, we show that the subnetwork-based QMLE is consistent and asymptotically normal under appropriate regularity conditions. Extensive simulation studies, based on both simulated and real network structures, are presented.

stat.CO↗

arXiv2023

Knowledge-Enhanced Relation Extraction Dataset

Recently, knowledge-enhanced methods leveraging auxiliary knowledge graphs have emerged in relation extraction, surpassing traditional text-based approaches. However, to our best knowledge, there is currently no public dataset available that encompasses both evidence sentences and knowledge graphs for knowledge-enhanced relation extraction. To address this gap, we introduce the Knowledge-Enhanced Relation Extraction Dataset (KERED). KERED annotates each sentence with a relational fact, and it provides knowledge context for entities through entity linking. Using our curated dataset, We compared contemporary relation extraction methods under two prevalent task settings: sentence-level and bag-level. The experimental result shows the knowledge graphs provided by KERED can support knowledge-enhanced relation extraction methods. We believe that KERED offers high-quality relation extraction datasets with corresponding knowledge graphs for evaluating the performance of knowledge-enhanced relation extraction methods. Our dataset is available at: \url{https://figshare.com/projects/KERED/134459}

cs.LG↗

arXiv2023

Statistical Analysis of Fixed Mini-Batch Gradient Descent Estimator

We study here a fixed mini-batch gradient decent (FMGD) algorithm to solve optimization problems with massive datasets. In FMGD, the whole sample is split into multiple non-overlapping partitions. Once the partitions are formed, they are then fixed throughout the rest of the algorithm. For convenience, we refer to the fixed partitions as fixed mini-batches. Then for each computation iteration, the gradients are sequentially calculated on each fixed mini-batch. Because the size of fixed mini-batches is typically much smaller than the whole sample size, it can be easily computed. This leads to much reduced computation cost for each computational iteration. It makes FMGD computationally efficient and practically more feasible. To demonstrate the theoretical properties of FMGD, we start with a linear regression model with a constant learning rate. We study its numerical convergence and statistical efficiency properties. We find that sufficiently small learning rates are necessarily required for both numerical convergence and statistical efficiency. Nevertheless, an extremely small learning rate might lead to painfully slow numerical convergence. To solve the problem, a diminishing learning rate scheduling strategy can be used. This leads to the FMGD estimator with faster numerical convergence and better statistical efficiency. Finally, the FMGD algorithms with random shuffling and a general loss function are also studied.

stat.CO↗

arXiv2023

Improved Naive Bayes with Mislabeled Data

Labeling mistakes are frequently encountered in real-world applications. If not treated well, the labeling mistakes can deteriorate the classification performances of a model seriously. To address this issue, we propose an improved Naive Bayes method for text classification. It is analytically simple and free of subjective judgements on the correct and incorrect labels. By specifying the generating mechanism of incorrect labels, we optimize the corresponding log-likelihood function iteratively by using an EM algorithm. Our simulation and experiment results show that the improved Naive Bayes method greatly improves the performances of the Naive Bayes method with mislabeled data.

cs.LG↗

arXiv2023

Testing Sufficiency for Transfer Learning

Modern statistical analysis often encounters high dimensional models but with limited sample sizes. This makes the target data based statistical estimation very difficult. Then how to borrow information from another large sized source data for more accurate target model estimation becomes an interesting problem. This leads to the useful idea of transfer learning. Various estimation methods in this regard have been developed recently. In this work, we study transfer learning from a different perspective. Specifically, we consider here the problem of testing for transfer learning sufficiency. By transfer learning sufficiency (denoted as the null hypothesis), we mean that, with the help of the source data, the useful information contained in the feature vectors of the target data can be sufficiently extracted for predicting the interested target response. Therefore, the rejection of the null hypothesis implies that information useful for prediction remains in the feature vectors of the target data and thus calls for further exploration. To this end, we develop a novel testing procedure and a centralized and standardized test statistic, whose asymptotic null distribution is analytically derived. Simulation studies are presented to demonstrate the finite sample performance of the proposed method. A deep learning related real data example is presented for illustration purpose.

stat.ME↗