Author name cluster

Jun Shen

Possible papers associated with this exact author name in Arrow. This page groups case-insensitive exact name matches and is not a full identity disambiguation profile.

13 papers

1 author row

EAAI Journal 2026 Journal Article

Enhanced graph neural network for rapid multi-field seismic prediction in shield tunnels with contact loss defects

Xianlong Wu
Jun Shen
Xiaohua Bao
Xiangsheng Chen
Hongzhi Cui

Contact loss defects (CLDs) frequently occur between tunnel linings and surrounding soil, substantially affecting soil–structure interaction and seismic behavior. Traditional finite element method (FEM) analyses are limited by complex modeling and high computational demands, making them impractical for large-scale or multi-scenario evaluations. To address these challenges, this study develops a graph neural network (GNN)-based framework to predict the multi-physics seismic response of shield tunnels with contact loss defects. The framework maps actual inspection data into a training dataset, using CLD parameters identified via ground-penetrating radar (GPR) and shear wave velocity as input features. A hybrid architecture combining multilayer perceptron (MLP) and GNN is employed to simultaneously predict radial displacement, Mises stress, and damage field distributions. Applied to a real-world shield tunnel project, the model achieved high prediction accuracy (R2 = 0. 98 for displacement, 0. 95 for stress, and 0. 92 for damage), with a total loss of 5. 4. Each prediction takes just 0. 15 s-over 5800 times faster than FEM simulations. To support practical use, the method has been implemented in an interactive tool, CLD-QuakePredictor V1. 0, demonstrating strong potential for efficient and scalable seismic performance assessment of shield tunnels.

Details DOI

AAAI Conference 2026 Conference Paper

Incomplete Multi-View Unsupervised Federated Feature Selection via Cooperative Particle Swarm Optimization and Tensor-Aligned Learning

Zhiwei Ye
Songsong Zhang
Wen Zhou
Libing Wu
Jun Shen
Ting Cai
Mingwei Wang
Jixin Zhang

With the widespread adoption of multi-view data in numerous fields, multi-view unsupervised feature selection (MUFS) has made notable strides in both feature pruning and missing-view completion. Nonetheless, existing MUFS methods typically rely on centralized servers, which cannot meet real-world demands for privacy preservation and distributed learning, and they often suffer from suboptimal solution and weak convergence guarantees. To address these challenges, IMUFFS, an incomplete multi-view unsupervised federated feature selection via cooperative particle swarm optimization (CPSO) and tensor-aligned learning (TAL) is proposed. Specifically, each client executes CPSO-TAL at two stages: (i) an external optimization phase that involves a CPSO, inspired by the co-evolutionary mechanism of hybrid breeding optimization algorithm, performing a global search in the feature space, and (ii) an internal optimization phase that leverages TAL with imputation and CP decomposition, where CP decomposition reduces dimensionality by decomposing the original tensor into a sum of core components, to learn low-dimensional embeddings, while simultaneously updating anchor graphs and view preference weights, thereby harmonizing imputation and representation learning. On the server side, a federated aggregation strategy using adaptive normalized mutual information (NMI) weighting combines the locally optimized feature selection (FS) weights and NMI scores from clients, ensuring privacy while improving the quality of FS and convergence. Extensive experiments on multiple datasets demonstrate that IMUFFS consistently outperforms state-of-the-art methods, yielding more effective and robust FS and enhancing better missing-view completion.

PDF Details DOI

AAAI Conference 2026 Conference Paper

Towards Zero-Shot Diabetic Retinopathy Grading: Learning Generalized Knowledge via Prompt-Driven Matching and Emulating

Huan Wang
Haoran Li
Yuxin Lin
Huaming Chen
Jun Yan
Lijuan Wang
Jiahua Shi
Qihao Xu

As one of the primary causes of visual impairment, Diabetic Retinopathy (DR) requires accurate and robust grading to facilitate timely diagnosis and intervention. Different from conventional DR grading methods that utilize single-view images, recent clinical studies have revealed that multi-view fundus images can significantly enhance DR grading performance by expanding the field of view (FOV). However, there is a long-tailed distribution problem in fundus image analysis, i.e., a high prevalence of mild DR grades and a low prevalence of rare ones (e.g., cases of high severity), which presents a significant challenge to developing a unified model capable of detecting rare or unseen DR grades not encountered during training. In this paper, we propose ProME-DR, a Prompt-driven zero-shot DR grading framework, which leverages prompt Matching and Emulating to recognize the unseen DR categories and views beyond the training set. ProME-DR disentangles the training process into two stages to learn generalized knowledge for novel DR disease grading. Initially, ProME-DR leverages two sets of prompt units to capture semantic and inter-view consistency knowledge via a split-and-mask manner, gathering instance-level DR visual clues. Subsequently, it constructs a concept-aware emulator to generate context prompt units, linking extensible knowledge learned from the previously seen DR attributes for zero-shot DR grading. Extensive experiments conducted on eight datasets and various scenarios confirm the superiority of ProME-DR.

PDF Details DOI

NeurIPS Conference 2025 Conference Paper

InstructSAM: A Training-free Framework for Instruction-Oriented Remote Sensing Object Recognition

Yijie Zheng
Weijie Wu
Qingyun Li
Xuehui Wang
Xu Zhou
Aiai Ren
Jun Shen
Long Zhao

Language-guided object recognition in remote sensing imagery is crucial for large-scale mapping and automated data annotation. However, existing open-vocabulary and visual grounding methods rely on explicit category cues, limiting their ability to handle complex or implicit queries that require advanced reasoning. To address this issue, we introduce a new suite of tasks, including Instruction-Oriented Object Counting, Detection, and Segmentation (InstructCDS), covering open-vocabulary, open-ended, and open-subclass scenarios. We further present EarthInstruct, the first InstructCDS benchmark for earth observation. It is constructed from two diverse remote sensing datasets with varying spatial resolutions and annotation rules across 20 categories, necessitating models to interpret dataset-specific instructions. Given the scarcity of semantically rich labeled data in remote sensing, we propose InstructSAM, a training-free framework for instruction-driven object recognition. InstructSAM leverages large vision-language models to interpret user instructions and estimate object counts, employs SAM2 for mask proposal, and formulates mask-label assignment as a binary integer programming problem. By integrating semantic similarity with counting constraints, InstructSAM efficiently assigns categories to predicted masks without relying on confidence thresholds. Experiments demonstrate that InstructSAM matches or surpasses specialized baselines across multiple tasks while maintaining near-constant inference time regardless of object count, reducing output tokens by 89\% and overall runtime by over 32\% compared to direct generation approaches. We believe the contributions of the proposed tasks, benchmark, and effective approach will advance future research in developing versatile object recognition systems. The code is available at https: //VoyagerXvoyagerx. github. io/InstructSAM.

PDF Details

IS Journal 2025 Journal Article

Machine Learning Approaches for Micromobility User Behavior Analysis

Cheng Zhang
Bo Du
Qiuyun Luan
Jun Shen

With widespread adoption globally, micromobility like bikes, e-scooters, and e-bikes has attracted increasing attention due to its ability to complement existing transportation modes and promote sustainable transportation. Understanding micromobility user behaviors in urban areas is essential for improving safety and comfort, as well as for informing infrastructure development and policy. Prior investigations on micromobility user behaviors primarily relied on statistical and kinematic modeling approaches. Although these methods have proven effective in characterizing user behaviors at both macroscopic and microscopic levels, the advent of artificial intelligence (AI)-powered data analytics and behavioral modeling is revolutionizing the field. Recently, advanced machine learning models, such as gradient boosting decision tree, graph convolutional network, and inverse reinforcement learning, has introduced new momentum into micromobility user behavior research. This article explores recent developments, research opportunities, and future directions in this field, leveraging the power of more generic AI approaches.

Details DOI

TIST Journal 2025 Journal Article

Personalized Learning Path Recommendation with Time-Aware Attention-Based Reinforcement Learning

Shantao Jiang
Yiping Wen
Jun Shen
Gaoxian Peng
Guosheng Kang
Jianxun Liu

Learning resources in online learning systems typically adhere to uniform formats and settings, lacking flexibility and personalization to meet diverse learning needs and preferences. This inability to meet individualized learning needs and preferences has spurred research interest in personalized learning path recommendations. Many researchers have explored recommending learning path by leveraging user historical learning resource sequence to model personalized characteristics. However, these methods overlook the time information in the learning process and fail to interpret the dynamic shifts in learning preferences during recommendation. Therefore, we propose a method, termed TA-RL, for learning path recommendation, based on time-aware attention mechanism and reinforcement learning. First, we propose a novel time-aware attention mechanism to trace the evolving learning preferences of user, in which attention weights are computed using a context-aware time distance measure and the similarity between history learning resources. Then, we employ a Monte Carlo policy gradient reinforcement learning method to generate learning path recommendation based on learning preferences. We validate the effectiveness of our proposed method by comprehensive experiments on two real-world datasets.

Details DOI

EAAI Journal 2025 Journal Article

Risk-response coupling in underground structures under liquefiable soil conditions: A causality-informed Dynamic Bayesian network integrated framework

Heqi Kong
Xiaohua Bao
Jun Shen
Xiangcou Zheng
Xiangsheng Chen

Underground structures in liquefiable soils face complex seismic risks that can trigger cascading failures. This study proposes a Granger causality-informed Dynamic Bayesian network (G-DBN) framework to capture the temporal propagation of seismic risk in such systems. Firstly, a system risk assessment model integrates multiple performance indicators through Cloud Model (CM) to quantify overall risk levels, considering uncertainties associated with soil liquefaction and structural responses. Subsequently, a structural dynamic risk inference model is established using Dynamic Bayesian network (DBN), combining Granger Causality (GC) analysis with engineering-informed relationships to define the network structure. The input features include key structural state variables such as tunnel cross-section convergence (T 4), tunnel uplift displacement (T 6), station uplift displacement (S 5), and inter-story drift angle (S 6), and the aggregated structural risk indicator serves as the target variable. This framework enables the temporal propagation of risk across interconnected structural nodes, and elucidates the mechanisms by which liquefiable soil deformations and structural responses interact within the soil-structure system. Results showed that the risk characteristic value (Expectation, E x) decreased from 29. 12 % (percentage) to 5. 21 % as the Peak Ground Acceleration (PGA, expressed in units of gravitational acceleration g) increased from 0. 1 g to 0. 7 g. The proposed G-DBN model demonstrates robust predictive capabilities, achieving coefficient of determination (R 2) values exceeding 0. 95 across multiple seismic intensity conditions. Additionally, tunnel cross-section convergence (T 4) was identified as the most critical factor affecting risk propagation in the coupled underground systems. By integrating holistic risk quantification with dynamic propagation analysis, this study offers a robust tool for understanding dynamic risk evolution and supports decision-making for seismic resilience of underground infrastructure in liquefiable soils.

Details DOI

AAAI Conference 2025 Conference Paper

Vox-UDA: Voxel-wise Unsupervised Domain Adaptation for Cryo-Electron Subtomogram Segmentation with Denoised Pseudo-Labeling

Haoran Li
Xingjian Li
Jiahua Shi
Huaming Chen
Bo Du
Daisuke Kihara
Johan Barthelemy
Jun Shen

Cryo-Electron Tomography (cryo-ET) is a 3D imaging technology that facilitates the study of macromolecular structures at near-atomic resolution. Recent volumetric segmentation approaches on cryo-ET images have drawn widespread interest in the biological sector. However, existing methods heavily rely on manually labeled data, which requires highly professional skills, thereby hindering the adoption of fully-supervised approaches for cryo-ET images. Some unsupervised domain adaptation (UDA) approaches have been designed to enhance the segmentation network performance using unlabeled data. However, applying these methods directly to cryo-ET image segmentation tasks remains challenging due to two main issues: 1) the source dataset, usually obtained through simulation, contains a fixed level of noise, while the target dataset, directly collected from raw-data from the real-world scenario, have unpredictable noise levels. 2) the source data used for training typically consists of known macromoleculars. In contrast, the target domain data are often unknown, causing the model to be biased towards those known macromolecules, leading to a domain shift problem. To address such challenges, in this work, we introduce a voxel-wise unsupervised domain adaptation approach, termed Vox-UDA, specifically for cryo-ET subtomogram segmentation. Vox-UDA incorporates a noise generation module to simulate target-like noises in the source dataset for cross-noise level adaptation. Additionally, we propose a denoised pseudo-labeling strategy based on the improved Bilateral Filter to alleviate the domain shift problem. More importantly, we construct the first UDA cryo-ET subtomogram segmentation benchmark on three experimental datasets. Extensive experimental results on multiple benchmarks and newly curated real-world datasets demonstrate the superiority of our proposed approach compared to state-of-the-art UDA methods.

PDF Details DOI

NeurIPS Conference 2024 Conference Paper

Metric from Human: Zero-shot Monocular Metric Depth Estimation via Test-time Adaptation

Yizhou Zhao
Hengwei Bian
Kaihua Chen
Pengliang Ji
Liao Qu
Shao-yu Lin
Weichen Yu
Haoran Li

Monocular depth estimation (MDE) is fundamental for deriving 3D scene structures from 2D images. While state-of-the-art monocular relative depth estimation (MRDE) excels in estimating relative depths for in-the-wild images, current monocular metric depth estimation (MMDE) approaches still face challenges in handling unseen scenes. Since MMDE can be viewed as the composition of MRDE and metric scale recovery, we attribute this difficulty to scene dependency, where MMDE models rely on scenes observed during supervised training for predicting scene scales during inference. To address this issue, we propose to use humans as landmarks for distilling scene-independent metric scale priors from generative painting models. Our approach, Metric from Human (MfH), bridges from generalizable MRDE to zero-shot MMDE in a generate-and-estimate manner. Specifically, MfH generates humans on the input image with generative painting and estimates human dimensions with an off-the-shelf human mesh recovery (HMR) model. Based on MRDE predictions, it propagates the metric information from painted humans to the contexts, resulting in metric depth estimations for the original input. Through this annotation-free test-time adaptation, MfH achieves superior zero-shot performance in MMDE, demonstrating its strong generalization ability.

PDF Details DOI

AAAI Conference 2018 Conference Paper

Splitting an LPMLN Program

Bin Wang
Zhizheng Zhang
Hongxiang Xu
Jun Shen

The technique called splitting sets has been proven useful in simplifying the investigation of Answer Set Programming (ASP). In this paper, we investigate the splitting set theorem for LPMLN that is a new extension of ASP created by combining the ideas of ASP and Markov Logic Networks (MLN). Firstly, we extend the notion of splitting sets to LPMLN programs and present the splitting set theorem for LPMLN. Then, the use of the theorem for simplifying several LPMLN inference tasks is illustrated. After that, we give two parallel approaches for solving LPMLN programs via using the theorem. The preliminary experimental results show that these approaches are alternative ways to promote an LPMLN solver.

PDF Details

YNIMG Journal 2006 Journal Article

In vivo detection of gray and white matter differences in GABA concentration in the human brain

In-Young Choi
Sang-Pil Lee
Hellmut Merkle
Jun Shen

A novel selective multiple quantum filtering-based chemical shift imaging method was developed for acquiring GABA images in the human brain at 3 T. This method allows a concomitant acquisition of an interleaved total creatine image with the same spatial resolution. Using T 1-based image segmentation and a nonlinear least square regression analysis of GABA-to-total creatine concentration ratios in frontal and parietal lobes of healthy adult volunteers as a function of the tissue gray matter fraction, the mean GABA concentration in gray and white matter was determined to be 1. 30±0. 36 μmol/g and 0. 16±0. 16 μmol/g (mean±SD, n =13), respectively. It is expected that this method will become a useful tool for studying GABAergic function in the human brain in vivo.

Details DOI

YNIMG Journal 2006 Journal Article

Increased oxygen consumption in the somatosensory cortex of α-chloralose anesthetized rats during forepaw stimulation determined using MRS at 11.7 Tesla

Jehoon Yang
Jun Shen

The significance of changes in cerebral oxygen consumption in focally activated brain tissue is still controversial. Since the rate of cerebral oxygen consumption is tightly coupled to that of tricarboxylic acid cycle which can be measured from the turnover kinetics of [4-13C]glutamate using in vivo 1H{13C} magnetic resonance spectroscopy, changes in tricarboxylic acid cycle flux rate were assessed in primary somatosensory cortex of α-chloralose anesthetized rats during electrical forepaw stimulation. With markedly improved 1H{13C} magnetic resonance spectroscopy technique and the use of high magnetic field strength of 11. 7 T accessible to the current study, [4-13C]glutamate at 2. 35 ppm was spectrally resolved from overlapping resonances of [4-13C]glutamine at 2. 46 ppm and [2-13C]GABA at 2. 28 ppm as well as the more distal [3-13C]glutamate and [3-13C]glutamine. The results showed a significantly increased V TCA in focally activated primary somatosensory cortex during forepaw stimulation, corresponding to approximately 51 ± 27% (n = 6, mean ± SD) increase in cerebral oxygen consumption rate. Considering the high efficiency in producing adenosine triphosphate by oxidative metabolism of glucose, the results demonstrate that aerobic oxidative metabolism provides the majority of energy required for cerebral focal activation in α-chloralose anesthetized rats subjected to forepaw stimulation.

Details DOI

YNIMG Journal 2005 Journal Article

Metabolic alterations in focally activated primary somatosensory cortex of α-chloralose-anesthetized rats measured by 1H MRS at 11.7 T

Su Xu
Jehoon Yang
Charles Q. Li
Wenjun Zhu
Jun Shen

Previously, magnetic resonance spectroscopy studies of alterations in cerebral metabolite concentration during functional activation have been focused on phosphocreatine using 31P MRS and lactate using 1H MRS with controversial results. Recently, significant improvements on the spectral resolution and sensitivity of in vivo spectroscopy have been made at ultrahigh magnetic field strength. Using highly resolved localized short-TE 1H MRS at 11. 7 T, we report metabolic responses of rat somatosensory cortex to forepaw stimulation in α-chloralose-anesthetized rats. The phosphocreatine/creatine ratio was found to be significantly decreased by 15. 1 ± 4. 6% (mean ± SEM, P < 0. 01). Lactate remained very low (∼<0. 3 μmol/g w/w) with no statistically significant changes observed during forepaw stimulation at a temporal resolution of 10. 7 min. An increase in glutamine and a decrease in glutamate and myo-inositol were also detected in the stimulated state. Our results suggest that, under the experimental conditions used in this study, increased energy consumption due to focal activation causes a shift in the creatine kinase reaction towards the direction of adenosine triphosphate production. At the same time, metabolic matching prevails during increased energy consumption with no significant increase in the glycolytic product lactate in the focally activated primary somatosensory cortex of α-chloralose-anesthetized rats.

Details DOI