arXiv subjects

Jia Wei

Publications and source records attributed to Jia Wei.

At least 37 records · Page 2Linked to original sources

arXiv2025

AWEMixer: Adaptive Wavelet-Enhanced Mixer Network for Long-Term Time Series Forecasting

Forecasting long-term time series in IoT environments remains a significant challenge due to the non-stationary and multi-scale characteristics of sensor signals. Furthermore, error accumulation causes a decrease in forecast quality when predicting further into the future. Traditional methods are restricted to operate in time-domain, while the global frequency information achieved by Fourier transform would be regarded as stationary signals leading to blur the temporal patterns of transient events. We propose AWEMixer, an Adaptive Wavelet-Enhanced Mixer Network including two innovative components: 1) a Frequency Router designs to utilize the global periodicity pattern achieved by Fast Fourier Transform to adaptively weight localized wavelet subband, and 2) a Coherent Gated Fusion Block to achieve selective integration of prominent frequency features with multi-scale temporal representation through cross-attention and gating mechanism, which realizes accurate time-frequency localization while remaining robust to noise. Seven public benchmarks validate that our model is more effective than recent state-of-the-art models. Specifically, our model consistently achieves performance improvement compared with transformer-based and MLP-based state-of-the-art models in long-sequence time series forecasting. Code is available at https://github.com/hit636/AWEMixer

cs.LG↗

arXiv2025

SageAttention: Accurate 8-Bit Attention for Plug-and-play Inference Acceleration

The transformer architecture predominates across various models. As the heart of the transformer, attention has a computational complexity of $O(N^2)$, compared to $O(N)$ for linear transformations. When handling large sequence lengths, attention becomes the primary time-consuming component. Although quantization has proven to be an effective method for accelerating model inference, existing quantization methods primarily focus on optimizing the linear layer. In response, we first analyze the feasibility of quantization in attention detailedly. Following that, we propose SageAttention, a highly efficient and accurate quantization method for attention. The OPS (operations per second) of our approach outperforms FlashAttention2 and xformers by about 2.1 times and 2.7 times, respectively. SageAttention also achieves superior accuracy performance over FlashAttention3. Comprehensive experiments confirm that our approach incurs almost no end-to-end metrics loss across diverse models, including those for large language processing, image generation, and video generation. The codes are available at https://github.com/thu-ml/SageAttention.

cs.LG↗

arXiv2025

SageAttention2: Efficient Attention with Thorough Outlier Smoothing and Per-thread INT4 Quantization

Although quantization for linear layers has been widely used, its application to accelerate the attention process remains limited. To further enhance the efficiency of attention computation compared to SageAttention while maintaining precision, we propose SageAttention2, which utilizes significantly faster 4-bit matrix multiplication (Matmul) alongside additional precision-enhancing techniques. First, we propose to quantize matrices $(Q, K)$ to INT4 in a hardware-friendly thread-level granularity and quantize matrices $(\widetilde P, V)$ to FP8. Second, we propose a method to smooth $Q$, enhancing the accuracy of INT4 $QK^\top$. Third, we propose a two-level accumulation strategy for $\widetilde PV$ to enhance the accuracy of FP8 $\widetilde PV$. The operations per second (OPS) of SageAttention2 surpass FlashAttention2 and xformers by about 3x and 4.5x on RTX4090, respectively. Moreover, SageAttention2 matches the speed of FlashAttention3(fp8) on the Hopper GPUs, while delivering much higher accuracy. Comprehensive experiments confirm that our approach incurs negligible end-to-end metrics loss across diverse models, including those for language, image, and video generation. The code is available at https://github.com/thu-ml/SageAttention.

cs.LG↗

arXiv2025

Multi-Modal Artificial Intelligence of Embryo Grading and Pregnancy Prediction in Assisted Reproductive Technology: A Review

Infertility, a pressing global health concern, affects a substantial proportion of individuals worldwide. While advancements in assisted reproductive technology (ART) have offered effective interventions, conventional in vitro fertilization-embryo transfer (IVF-ET) procedures still encounter significant hurdles in enhancing pregnancy success rates. Key challenges include the inherent subjectivity in embryo grading and the inefficiency of multi-modal data integration. Against this backdrop, the adoption of AI-driven technologies has emerged as a pivotal strategy to address these issues. This article presents a comprehensive review of the progress in AI applications for embryo grading and pregnancy prediction from a novel perspective, with a specific focus on the utilization of different modal data, such as static images, time-lapse videos, and structured tabular data. The reason for this perspective is that reorganizing tasks based on data sources can not only more accurately depict the essence of the problem but also help clarify the rationality and limitations of model design. Furthermore, this review critically examines the core challenges in contemporary research, encompassing the intricacies of multi-modal feature fusion, constraints imposed by data scarcity, limitations in model generalization capabilities, and the dynamically evolving legal and regulatory frameworks. On this basis, it explicitly identifies potential avenues for future research, aiming to provide actionable guidance for advancing the application of multi-modal AI in the field of ART.

cs.AI↗

arXiv2025

DPANet: Dual Pyramid Attention Network for Multivariate Time Series Forecasting

Long-term time series forecasting (LTSF) is hampered by the challenge of modeling complex dependencies that span multiple temporal scales and frequency resolutions. Existing methods, including Transformer and MLP-based models, often struggle to capture these intertwined characteristics in a unified and structured manner. We propose the Dual Pyramid Attention Network (DPANet), a novel architecture that explicitly decouples and concurrently models temporal multi-scale dynamics and spectral multi-resolution periodicities. DPANet constructs two parallel pyramids: a Temporal Pyramid built on progressive downsampling, and a Frequency Pyramid built on band-pass filtering. The core of our model is the Cross-Pyramid Fusion Block, which facilitates deep, interactive information exchange between corresponding pyramid levels via cross-attention. This fusion proceeds in a coarse-to-fine hierarchy, enabling global context to guide local representation learning. Extensive experiments on public benchmarks show that DPANet achieves state-of-the-art performance, significantly outperforming prior models. Code is available at https://github.com/hit636/DPANet.

cs.LG↗

arXiv2025

When FinTech Meets Privacy: Securing Financial LLMs with Differential Private Fine-Tuning

The integration of Large Language Models (LLMs) into financial technology (FinTech) has revolutionized the analysis and processing of complex financial data, driving advancements in real-time decision-making and analytics. With the growing trend of deploying AI models on edge devices for financial applications, ensuring the privacy of sensitive financial data has become a significant challenge. To address this, we propose DPFinLLM, a privacy-enhanced, lightweight LLM specifically designed for on-device financial applications. DPFinLLM combines a robust differential privacy mechanism with a streamlined architecture inspired by state-of-the-art models, enabling secure and efficient processing of financial data. This proposed DPFinLLM can not only safeguard user data from privacy breaches but also ensure high performance across diverse financial tasks. Extensive experiments on multiple financial sentiment datasets validate the effectiveness of DPFinLLM, demonstrating its ability to achieve performance comparable to fully fine-tuned models, even under strict privacy constraints.

cs.CR↗

arXiv2025

DeFusion: An Effective Decoupling Fusion Network for Multi-Modal Pregnancy Prediction

Temporal embryo images and parental fertility table indicators are both valuable for pregnancy prediction in \textbf{in vitro fertilization embryo transfer} (IVF-ET). However, current machine learning models cannot make full use of the complementary information between the two modalities to improve pregnancy prediction performance. In this paper, we propose a Decoupling Fusion Network called DeFusion to effectively integrate the multi-modal information for IVF-ET pregnancy prediction. Specifically, we propose a decoupling fusion module that decouples the information from the different modalities into related and unrelated information, thereby achieving a more delicate fusion. And we fuse temporal embryo images with a spatial-temporal position encoding, and extract fertility table indicator information with a table transformer. To evaluate the effectiveness of our model, we use a new dataset including 4046 cases collected from Southern Medical University. The experiments show that our model outperforms state-of-the-art methods. Meanwhile, the performance on the eye disease prediction dataset reflects the model's good generalization. Our code is available at https://github.com/Ou-Young-1999/DFNet.

cs.CV↗

arXiv2025

Accurate INT8 Training Through Dynamic Block-Level Fallback

Transformer models have achieved remarkable success across various AI applications but face significant training costs. Low-bit training, such as INT8 training, can leverage computational units with higher throughput, and has already demonstrated its effectiveness on GPT2 models with block-level quantization. However, it struggles with modern Transformer variants incorporating GLU units. This is because those variants demonstrate complex distributions of activation outliers. To address the challenge, we propose Fallback Quantization, implementing mixed-precision GEMM that dynamically falls back 8-bit to 16-bit for activation blocks containing outliers. Experiments show that our approach is robustly competent in both fine-tuning and pretraining settings. Moreover, our method achieves a 1.57x end-to-end training speedup on RTX4090 GPUs.

cs.LG↗

arXiv2025

SageAttention2++: A More Efficient Implementation of SageAttention2

The efficiency of attention is critical because its time complexity grows quadratically with sequence length. SageAttention2 addresses this by utilizing quantization to accelerate matrix multiplications (Matmul) in attention. To further accelerate SageAttention2, we propose to utilize the faster instruction of FP8 Matmul accumulated in FP16. The instruction is 2x faster than the FP8 Matmul used in SageAttention2. Our experiments show that SageAttention2++ achieves a 3.9x speedup over FlashAttention while maintaining the same attention accuracy as SageAttention2. This means SageAttention2++ effectively accelerates various models, including those for language, image, and video generation, with negligible end-to-end metrics loss. The code will be available at https://github.com/thu-ml/SageAttention.

cs.LG↗

arXiv2025

Curation and Analysis of MIMICEL -- An Event Log for MIMIC-IV Emergency Department

The global issue of overcrowding in emergency departments (ED) necessitates the analysis of patient flow through ED to enhance efficiency and alleviate overcrowding. However, traditional analytical methods are time-consuming and costly. The healthcare industry is embracing process mining tools to analyse healthcare processes and patient flows. Process mining aims to discover, monitor, and enhance processes by obtaining knowledge from event log data. However, the availability of event logs is a prerequisite for applying process mining techniques. Hence, this paper aims to generate an event log for analysing processes in ED. In this study, we extract an event log from the MIMIC-IV-ED dataset and name it MIMICEL. MIMICEL captures the process of patient journey in ED, allowing for analysis of patient flows and improving ED efficiency. We present analyses conducted using MIMICEL to demonstrate the utility of the dataset. The curation of MIMICEL facilitates extensive use of MIMIC-IV-ED data for ED analysis using process mining techniques, while also providing the process mining research communities with a valuable dataset for study.

cs.DB↗

arXiv2025

Probing orbital magnetism of a kagome metal CsV3Sb5 by a tuning fork resonator

The recently discovered kagome metal CsV$_3$Sb$_5$ exhibits a complex phase diagram that encompasses frustrated magnetism, topological charge density wave (CDW), and superconductivity. One CDW state that breaks time-reversal symmetry was proposed in this compound, while the exact nature of the putative magnetic state remains elusive. To examine the thermodynamic state of CsV$_3$Sb$_5$ and assess the character of the associated magnetism, we conducted tuning fork resonator measurements of magnetotropic susceptibility over a broad range of angles, magnetic fields, and temperature. We found a cascade of phase transition in the CDW phase. Of particular interest is a highly anisotropic magnetic structure that arises below about 30~K, with a magnetic moment along the $c$-axis that has an extremely small magnitude. This magnetic state demonstrates extremely slow dynamics and small saturate field, all suggest that electronic phase below 30~K breaks time reversal symmetry and has an unconventional origin.

cond-mat.mtrl-sci↗

arXiv2025

Mixture-of-Shape-Experts (MoSE): End-to-End Shape Dictionary Framework to Prompt SAM for Generalizable Medical Segmentation

Single domain generalization (SDG) has recently attracted growing attention in medical image segmentation. One promising strategy for SDG is to leverage consistent semantic shape priors across different imaging protocols, scanner vendors, and clinical sites. However, existing dictionary learning methods that encode shape priors often suffer from limited representational power with a small set of offline computed shape elements, or overfitting when the dictionary size grows. Moreover, they are not readily compatible with large foundation models such as the Segment Anything Model (SAM). In this paper, we propose a novel Mixture-of-Shape-Experts (MoSE) framework that seamlessly integrates the idea of mixture-of-experts (MoE) training into dictionary learning to efficiently capture diverse and robust shape priors. Our method conceptualizes each dictionary atom as a shape expert, which specializes in encoding distinct semantic shape information. A gating network dynamically fuses these shape experts into a robust shape map, with sparse activation guided by SAM encoding to prevent overfitting. We further provide this shape map as a prompt to SAM, utilizing the powerful generalization capability of SAM through bidirectional integration. All modules, including the shape dictionary, are trained in an end-to-end manner. Extensive experiments on multiple public datasets demonstrate its effectiveness.

cs.CV↗

arXiv2025

Multi-Scale Diffusion: Enhancing Spatial Layout in High-Resolution Panoramic Image Generation

Diffusion models have recently gained recognition for generating diverse and high-quality content, especially in image synthesis. These models excel not only in creating fixed-size images but also in producing panoramic images. However, existing methods often struggle with spatial layout consistency when producing high-resolution panoramas due to the lack of guidance on the global image layout. This paper introduces the Multi-Scale Diffusion (MSD), an optimized framework that extends the panoramic image generation framework to multiple resolution levels. Our method leverages gradient descent techniques to incorporate structural information from low-resolution images into high-resolution outputs. Through comprehensive qualitative and quantitative evaluations against prior work, we demonstrate that our approach significantly improves the coherence of high-resolution panorama generation.

cs.CV↗

arXiv2024

$\textit{Dirigo}$: A Method to Extract Event Logs for Object-Centric Processes

Real-world processes involve multiple object types with intricate interrelationships. Traditional event logs (in XES format), which record process execution centred around the case notion, are restricted to a single-object perspective, making it difficult to capture the behaviour of multiple objects and their interactions. To address this limitation, object-centric event logs (OCEL) have been introduced to capture both the objects involved in a process and their interactions with events. The object-centric event data (OCED) metamodel extends the OCEL format by further capturing dynamic object attributes and object-to-object relations. Recently OCEL 2.0 has been proposed based on OCED metamodel. Current research on generating OCEL logs requires specific input data sources, and resulting log data often fails to fully conform to OCEL 2.0. Moreover, the generated OCEL logs vary across different representational formats and their quality remains unevaluated. To address these challenges, a set of quality criteria for evaluating OCEL log representations is established. Guided by these criteria, $\textit{Dirigo}$ is proposed -- a method for extracting event logs that not only conforms to OCEL 2.0 but also extends it by capturing the temporal aspect of dynamic object-to-object relations. Object-role Modelling (ORM), a conceptual data modelling technique, is employed to describe the artifact produced at each step of $\textit{Dirigo}$. To validate the applicability of $\textit{Dirigo}$, it is applied to a real-life use case, extracting an event log via simulation. The quality of the log representation of the extracted event log is compared to those of existing OCEL logs using the established quality criteria.

cs.DB↗

arXiv2024

SAM-Swin: SAM-Driven Dual-Swin Transformers with Adaptive Lesion Enhancement for Laryngo-Pharyngeal Tumor Detection

Laryngo-pharyngeal cancer (LPC) is a highly lethal malignancy in the head and neck region. Recent advancements in tumor detection, particularly through dual-branch network architectures, have significantly improved diagnostic accuracy by integrating global and local feature extraction. However, challenges remain in accurately localizing lesions and fully capitalizing on the complementary nature of features within these branches. To address these issues, we propose SAM-Swin, an innovative SAM-driven Dual-Swin Transformer for laryngo-pharyngeal tumor detection. This model leverages the robust segmentation capabilities of the Segment Anything Model 2 (SAM2) to achieve precise lesion segmentation. Meanwhile, we present a multi-scale lesion-aware enhancement module (MS-LAEM) designed to adaptively enhance the learning of nuanced complementary features across various scales, improving the quality of feature extraction and representation. Furthermore, we implement a multi-scale class-aware guidance (CAG) loss that delivers multi-scale targeted supervision, thereby enhancing the model's capacity to extract class-specific features. To validate our approach, we compiled three LPC datasets from the First Affiliated Hospital (FAHSYSU), the Sixth Affiliated Hospital (SAHSYSU) of Sun Yat-sen University, and Nanfang Hospital of Southern Medical University (NHSMU). The FAHSYSU dataset is utilized for internal training, while the SAHSYSU and NHSMU datasets serve for external evaluation. Extensive experiments demonstrate that SAM-Swin outperforms state-of-the-art methods, showcasing its potential for advancing LPC detection and improving patient outcomes. The source code of SAM-Swin is available at the URL of \href{https://github.com/VVJia/SAM-Swin}{https://github.com/VVJia/SAM-Swin}.

cs.CV↗

arXiv2024

SAM-FNet: SAM-Guided Fusion Network for Laryngo-Pharyngeal Tumor Detection

Laryngo-pharyngeal cancer (LPC) is a highly fatal malignant disease affecting the head and neck region. Previous studies on endoscopic tumor detection, particularly those leveraging dual-branch network architectures, have shown significant advancements in tumor detection. These studies highlight the potential of dual-branch networks in improving diagnostic accuracy by effectively integrating global and local (lesion) feature extraction. However, they are still limited in their capabilities to accurately locate the lesion region and capture the discriminative feature information between the global and local branches. To address these issues, we propose a novel SAM-guided fusion network (SAM-FNet), a dual-branch network for laryngo-pharyngeal tumor detection. By leveraging the powerful object segmentation capabilities of the Segment Anything Model (SAM), we introduce the SAM into the SAM-FNet to accurately segment the lesion region. Furthermore, we propose a GAN-like feature optimization (GFO) module to capture the discriminative features between the global and local branches, enhancing the fusion feature complementarity. Additionally, we collect two LPC datasets from the First Affiliated Hospital (FAHSYSU) and the Sixth Affiliated Hospital (SAHSYSU) of Sun Yat-sen University. The FAHSYSU dataset is used as the internal dataset for training the model, while the SAHSYSU dataset is used as the external dataset for evaluating the model's performance. Extensive experiments on both datasets of FAHSYSU and SAHSYSU demonstrate that the SAM-FNet can achieve competitive results, outperforming the state-of-the-art counterparts. The source code of SAM-FNet is available at the URL of https://github.com/VVJia/SAM-FNet.

cs.CV↗

arXiv2024

Unsupervised Tumor-Aware Distillation for Multi-Modal Brain Image Translation

Multi-modal brain images from MRI scans are widely used in clinical diagnosis to provide complementary information from different modalities. However, obtaining fully paired multi-modal images in practice is challenging due to various factors, such as time, cost, and artifacts, resulting in modality-missing brain images. To address this problem, unsupervised multi-modal brain image translation has been extensively studied. Existing methods suffer from the problem of brain tumor deformation during translation, as they fail to focus on the tumor areas when translating the whole images. In this paper, we propose an unsupervised tumor-aware distillation teacher-student network called UTAD-Net, which is capable of perceiving and translating tumor areas precisely. Specifically, our model consists of two parts: a teacher network and a student network. The teacher network learns an end-to-end mapping from source to target modality using unpaired images and corresponding tumor masks first. Then, the translation knowledge is distilled into the student network, enabling it to generate more realistic tumor areas and whole images without masks. Experiments show that our model achieves competitive performance on both quantitative and qualitative evaluations of image quality compared with state-of-the-art methods. Furthermore, we demonstrate the effectiveness of the generated images on downstream segmentation tasks. Our code is available at https://github.com/scut-HC/UTAD-Net.

eess.IV↗

arXiv2024

BEND: Bagging Deep Learning Training Based on Efficient Neural Network Diffusion

Bagging has achieved great success in the field of machine learning by integrating multiple base classifiers to build a single strong classifier to reduce model variance. The performance improvement of bagging mainly relies on the number and diversity of base classifiers. However, traditional deep learning model training methods are expensive to train individually and difficult to train multiple models with low similarity in a restricted dataset. Recently, diffusion models, which have been tremendously successful in the fields of imaging and vision, have been found to be effective in generating neural network model weights and biases with diversity. We creatively propose a Bagging deep learning training algorithm based on Efficient Neural network Diffusion (BEND). The originality of BEND comes from the first use of a neural network diffusion model to efficiently build base classifiers for bagging. Our approach is simple but effective, first using multiple trained model weights and biases as inputs to train autoencoder and latent diffusion model to realize a diffusion model from noise to valid neural network parameters. Subsequently, we generate several base classifiers using the trained diffusion model. Finally, we integrate these ba se classifiers for various inference tasks using the Bagging method. Resulting experiments on multiple models and datasets show that our proposed BEND algorithm can consistently outperform the mean and median accuracies of both the original trained model and the diffused model. At the same time, new models diffused using the diffusion model have higher diversity and lower cost than multiple models trained using traditional methods. The BEND approach successfully introduces diffusion models into the new deep learning training domain and provides a new paradigm for future deep learning training and inference.

cs.LG↗