arXiv subjects

Jinfeng Wang

Publications and source records attributed to Jinfeng Wang.

At least 19 recordsLinked to original sources

arXiv2026

Anomalous Orbital Reconstruction Controlled by Interorbital Correlations and Hund's Coupling

The microscopic origin of anomalous orbital polarization in a class of low-dimensional correlated oxides remains unresolved due to the competition among crystal-field effects, electronic correlations, and orbital-dependent dimensionality. Using dynamical mean-field theory, we investigate a two orbital Hubbard model with orbital-dependent dimensionality and reveal the mechanism for anomalous orbital polarization through orbital reconstruction beyond the bare crystal-field picture. We identify interorbital Coulomb interaction-induced charge competition as a key microscopic mechanism responsible for the orbital redistribution, which leads first to an orbital-polarized correlated metal and subsequently to an orbital-polarized Mott insulator. We further find that Hund's coupling acts as a filling-dependent regulator of orbital reconstruction. It weakens the correlation-induced orbital redistribution at quarter filling by competing with interorbital charge fluctuations, while at half filling it completely suppresses the orbital-polarized state by stabilizing high-spin orbital-balanced configurations. These results provide a unified picture of correlation driven orbital reconstruction and highlight the relevance of interorbital interactions and Hund's coupling for understanding orbital phenomena in low-dimensional transition-metal oxides.

cond-mat.str-el↗

arXiv2026

Navigating the Proximity-Safety Balance: Constraint Decomposition for Human Following in Pedestrian Crowds

Following a target human in crowded environments involves an inherent conflict between staying close to the target and navigating safely among surrounding pedestrians and obstacles. This conflict becomes more severe in dense scenarios, where aggressive following risks collisions and conservative margins lead to target loss, especially when pedestrian behaviors are unfamiliar or unpredictable. Existing reinforcement learning (RL) methods typically encode these competing objectives into a single dense reward, but the resulting proximity-safety balance is implicit and difficult to adjust across conditions. To address this, we decompose the human-following task into a sparse task reward and independent cost constraints within a multi-constraint RL formulation, where each constraint is managed through cost thresholds with direct behavioral meaning rather than implicit reward weight ratios, allowing explicit and tunable control over the trade-off. We further quantify the prediction uncertainty of human motions and integrate these estimates into the RL costs to enhance safety under unpredictable conditions. Extensive experiments across both in-distribution and out-of-distribution settings demonstrate that our method achieves an effective proximity-safety balance compared to baselines. Real-robot deployment further validates the feasibility of our method in real-world scenarios. More details are available on our project page: https://nav-ps-balance.github.io/.

cs.RO↗

arXiv2026

TAVR-VLM: Risk-Conditioned Causal Grounding for Hallucination-Resistant Report Generation

Transcatheter Aortic Valve Replacement (TAVR) planning requires meticulous multimodal reasoning. However, adapting Multimodal Large Language Models (MLLMs) to this high-stakes domain is severely impeded by diagnostic hallucinations, where generated text lacks anatomical grounding. To address this, TAVR-VLM is introduced: a novel framework featuring Risk-Conditioned Causal Grounding Attention (R-CGA) that instantiates a model-internal ``Risk $\rightarrow$ Region $\rightarrow$ Word'' structural grounding pathway. R-CGA compresses multimodal inputs into a causal risk bottleneck, purifying dense visual features into a global risk mask. During autoregressive generation, a support-projected causal consistency objective constrains token-level grounding within the risk-defined support mask. Evaluated on $\text{M}^3\text{TAVR}$, a comprehensive 1,482-patient cohort, TAVR-VLM establishes a new state-of-the-art. It achieves an AUROC of 0.896, boosts CIDEr to 0.936, and drastically reduces the hallucination rate to 8.1\%, thereby improving interpretability for evidence-based surgical AI.

cs.AI↗

arXiv2026

One-Shot Price Forecasting with Covariate-Guided Experts under Privacy Constraints

Forecasting in power systems often involves multivariate time series with complex dependencies and strict privacy constraints across regions. Traditional forecasting methods require significant expert knowledge and struggle to generalize across diverse deployment scenarios. Recent advancements in pre-trained time series models offer new opportunities, but their zero-shot performance on domain-specific tasks remains limited. To address these challenges, we propose a novel MoE Encoder module that augments pretrained forecasting models by injecting a sparse mixture-of-experts layer between tokenization and encoding. This design enables two key capabilities: (1) trans forming multivariate forecasting into an expert-guided univariate task, allowing the model to effectively capture inter-variable relations, and (2) supporting localized training and lightweight parameter sharing in federated settings where raw data cannot be exchanged. Extensive experiments on public multivariate datasets demonstrate that MoE-Encoder significantly improves forecasting accuracy compared to strong baselines. We further simulate federated environments and show that transferring only MoE-Encoder parameters allows efficient adaptation to new regions, with minimal performance degradation. Our findings suggest that MoE-Encoder provides a scalable and privacy-aware extension to foundation time series models.

cs.LG↗

arXiv2025

A novel Fourier Adjacency Transformer for advanced EEG emotion recognition

EEG emotion recognition faces significant hurdles due to noise interference, signal nonstationarity, and the inherent complexity of brain activity which make accurately emotion classification. In this study, we present the Fourier Adjacency Transformer, a novel framework that seamlessly integrates Fourier-based periodic analysis with graph-driven structural modeling. Our method first leverages novel Fourier-inspired modules to extract periodic features from embedded EEG signals, effectively decoupling them from aperiodic components. Subsequently, we employ an adjacency attention scheme to reinforce universal inter-channel correlation patterns, coupling these patterns with their sample-based counterparts. Empirical evaluations on SEED and DEAP datasets demonstrate that our method surpasses existing state-of-the-art techniques, achieving an improvement of approximately 6.5% in recognition accuracy. By unifying periodicity and structural insights, this framework offers a promising direction for future research in EEG emotion analysis.

eess.SP↗

arXiv2024

Contrastive Learning Via Equivariant Representation

Invariant Contrastive Learning (ICL) methods have achieved impressive performance across various domains. However, the absence of latent space representation for distortion (augmentation)-related information in the latent space makes ICL sub-optimal regarding training efficiency and robustness in downstream tasks. Recent studies suggest that introducing equivariance into Contrastive Learning (CL) can improve overall performance. In this paper, we revisit the roles of augmentation strategies and equivariance in improving CL's efficacy. We propose CLeVER (Contrastive Learning Via Equivariant Representation), a novel equivariant contrastive learning framework compatible with augmentation strategies of arbitrary complexity for various mainstream CL backbone models. Experimental results demonstrate that CLeVER effectively extracts and incorporates equivariant information from practical natural images, thereby improving the training efficiency and robustness of baseline models in downstream tasks and achieving state-of-the-art (SOTA) performance. Moreover, we find that leveraging equivariant information extracted by CLeVER simultaneously enhances rotational invariance and sensitivity across experimental tasks, and helps stabilize the framework when handling complex augmentations, particularly for models with small-scale backbones.

cs.LG↗

arXiv2024

DualStreamFoveaNet: A Dual Stream Fusion Architecture with Anatomical Awareness for Robust Fovea Localization

Accurate fovea localization is essential for analyzing retinal diseases to prevent irreversible vision loss. While current deep learning-based methods outperform traditional ones, they still face challenges such as the lack of local anatomical landmarks around the fovea, the inability to robustly handle diseased retinal images, and the variations in image conditions. In this paper, we propose a novel transformer-based architecture called DualStreamFoveaNet (DSFN) for multi-cue fusion. This architecture explicitly incorporates long-range connections and global features using retina and vessel distributions for robust fovea localization. We introduce a spatial attention mechanism in the dual-stream encoder to extract and fuse self-learned anatomical information, focusing more on features distributed along blood vessels and significantly reducing computational costs by decreasing token numbers. Our extensive experiments show that the proposed architecture achieves state-of-the-art performance on two public datasets and one large-scale private dataset. Furthermore, we demonstrate that the DSFN is more robust on both normal and diseased retina images and has better generalization capacity in cross-dataset experiments.

cs.CV↗

arXiv2024

ProMISe: Promptable Medical Image Segmentation using SAM

With the proposal of the Segment Anything Model (SAM), fine-tuning SAM for medical image segmentation (MIS) has become popular. However, due to the large size of the SAM model and the significant domain gap between natural and medical images, fine-tuning-based strategies are costly with potential risk of instability, feature damage and catastrophic forgetting. Furthermore, some methods of transferring SAM to a domain-specific MIS through fine-tuning strategies disable the model's prompting capability, severely limiting its utilization scenarios. In this paper, we propose an Auto-Prompting Module (APM), which provides SAM-based foundation model with Euclidean adaptive prompts in the target domain. Our experiments demonstrate that such adaptive prompts significantly improve SAM's non-fine-tuned performance in MIS. In addition, we propose a novel non-invasive method called Incremental Pattern Shifting (IPS) to adapt SAM to specific medical domains. Experimental results show that the IPS enables SAM to achieve state-of-the-art or competitive performance in MIS without the need for fine-tuning. By coupling these two methods, we propose ProMISe, an end-to-end non-fine-tuned framework for Promptable Medical Image Segmentation. Our experiments demonstrate that both using our methods individually or in combination achieves satisfactory performance in low-cost pattern shifting, with all of SAM's parameters frozen.

cs.CV↗

arXiv2024

Geotree of Geodetector: An Anatomy of Knowledge Diffusion of a Novel Statistic

The growing number of citations to original publications highlighted their utility across academia, but the dissemination of knowledge from tacit conceptualization to scientific publications and its global applications remains understudied, and the prediction of knowledge trends in a disciplinary context is rare. Addressing the gaps, this paper constructed a tree-like hierarchical model (Geotree) to dissect the knowledge evolution paths of the Geodetector theory (a case) using the Web of Science citation database. Our results revealed that the knowledge evolution of 932 citations to Geodetector was partitioned into periods: a budding period of initial theoretical exploration, a growing period for emerging topics in application, and a mature period marked by significant citation growth. Our test R2 of the predicting model over the next decade, considering the tree-like hierarchy across research directions and disciplines, was 100% higher than that of the other two (from 0.29 to 0.58). The knowledge spreading, from China to North America in 2011, Europe in 2012, Oceania in 2017, South America in 2018, and Africa in 2019, was more associated with a country s production of scientific publications (q-statistic = 0.307***) than its income level. The Geotree modeling of two other cases from space science and physics confirmed the reliability of the source publication-based approach in tracking knowledge diffusion. Our established research framework enriched the current methodology of information science and provided valuable references for policymakers and scholars to enhance their decision-making processes.

cs.DL↗

arXiv2023

Distortion-Disentangled Contrastive Learning

Self-supervised learning is well known for its remarkable performance in representation learning and various downstream computer vision tasks. Recently, Positive-pair-Only Contrastive Learning (POCL) has achieved reliable performance without the need to construct positive-negative training sets. It reduces memory requirements by lessening the dependency on the batch size. The POCL method typically uses a single loss function to extract the distortion invariant representation (DIR) which describes the proximity of positive-pair representations affected by different distortions. This loss function implicitly enables the model to filter out or ignore the distortion variant representation (DVR) affected by different distortions. However, existing POCL methods do not explicitly enforce the disentanglement and exploitation of the actually valuable DVR. In addition, these POCL methods have been observed to be sensitive to augmentation strategies. To address these limitations, we propose a novel POCL framework named Distortion-Disentangled Contrastive Learning (DDCL) and a Distortion-Disentangled Loss (DDL). Our approach is the first to explicitly disentangle and exploit the DVR inside the model and feature stream to improve the overall representation utilization efficiency, robustness and representation ability. Experiments carried out demonstrate the superiority of our framework to Barlow Twins and Simsiam in terms of convergence, representation quality, and robustness on several benchmark datasets.

cs.CV↗

arXiv2023

Higher-order topological and nodal superconducting transition-metal sulfides MS (M = Nb and Ta)

Intrinsic topological superconducting materials are exotic and vital to develop the next-generation topological superconducting devices, topological quantum calculations, and quantum information technologies. Here, we predict the topological and nodal superconductivity of MS (M = Nb and Ta) transition-metal sulfides by using the density functional theory for superconductors combining with the symmetry indicators. We reveal their higher-order topology nature with an index of Z4 = 2. These materials have a higher Tc than the Nb or Ta metal superconductors due to their flat-band and strong electron-phonon coupling nature. Electron doping and lighter isotopes can effectively enhance the Tc. Our findings show that the MS (M = Nb and Ta) systems can be new platforms to study exotic physics in the higher-order topological superconductors, and provide a theoretical support to utilize them as the topological superconducting devices in the field of advanced topological quantum calculations and information technologies.

cond-mat.supr-con↗

arXiv2023

Atrial Septal Defect Detection in Children Based on Ultrasound Video Using Multiple Instances Learning

Purpose: Congenital heart defect (CHD) is the most common birth defect. Thoracic echocardiography (TTE) can provide sufficient cardiac structure information, evaluate hemodynamics and cardiac function, and is an effective method for atrial septal defect (ASD) examination. This paper aims to study a deep learning method based on cardiac ultrasound video to assist in ASD diagnosis. Materials and methods: We select two standard views of the atrial septum (subAS) and low parasternal four-compartment view (LPS4C) as the two views to identify ASD. We enlist data from 300 children patients as part of a double-blind experiment for five-fold cross-validation to verify the performance of our model. In addition, data from 30 children patients (15 positives and 15 negatives) are collected for clinician testing and compared to our model test results (these 30 samples do not participate in model training). We propose an echocardiography video-based atrial septal defect diagnosis system. In our model, we present a block random selection, maximal agreement decision and frame sampling strategy for training and testing respectively, resNet18 and r3D networks are used to extract the frame features and aggregate them to build a rich video-level representation. Results: We validate our model using our private dataset by five-cross validation. For ASD detection, we achieve 89.33 AUC, 84.95 accuracy, 85.70 sensitivity, 81.51 specificity and 81.99 F1 score. Conclusion: The proposed model is multiple instances learning-based deep learning model for video atrial septal defect detection which effectively improves ASD detection accuracy when compared to the performances of previous networks and clinical doctors.

eess.IV↗

arXiv2023

A Robust Framework of Chromosome Straightening with ViT-Patch GAN

Chromosomes carry the genetic information of humans. They exhibit non-rigid and non-articulated nature with varying degrees of curvature. Chromosome straightening is an important step for subsequent karyotype construction, pathological diagnosis and cytogenetic map development. However, robust chromosome straightening remains challenging, due to the unavailability of training images, distorted chromosome details and shapes after straightening, as well as poor generalization capability. In this paper, we propose a novel architecture, ViT-Patch GAN, consisting of a self-learned motion transformation generator and a Vision Transformer-based patch (ViT-Patch) discriminator. The generator learns the motion representation of chromosomes for straightening. With the help of the ViT-Patch discriminator, the straightened chromosomes retain more shape and banding pattern details. The experimental results show that the proposed method achieves better performance on Fréchet Inception Distance (FID), Learned Perceptual Image Patch Similarity (LPIPS) and downstream chromosome classification accuracy, and shows excellent generalization capability on a large dataset.

cs.CV↗

arXiv2022

DuAT: Dual-Aggregation Transformer Network for Medical Image Segmentation

Transformer-based models have been widely demonstrated to be successful in computer vision tasks by modelling long-range dependencies and capturing global representations. However, they are often dominated by features of large patterns leading to the loss of local details (e.g., boundaries and small objects), which are critical in medical image segmentation. To alleviate this problem, we propose a Dual-Aggregation Transformer Network called DuAT, which is characterized by two innovative designs, namely, the Global-to-Local Spatial Aggregation (GLSA) and Selective Boundary Aggregation (SBA) modules. The GLSA has the ability to aggregate and represent both global and local spatial features, which are beneficial for locating large and small objects, respectively. The SBA module is used to aggregate the boundary characteristic from low-level features and semantic information from high-level features for better preserving boundary details and locating the re-calibration objects. Extensive experiments in six benchmark datasets demonstrate that our proposed model outperforms state-of-the-art methods in the segmentation of skin lesion images, and polyps in colonoscopy images. In addition, our approach is more robust than existing methods in various challenging situations such as small object segmentation and ambiguous object boundaries.

cs.CV↗

arXiv2022

Statistics for Spatially Stratified Heterogeneous Data

Spatial statistics is dominated by spatial autocorrelation (SAC) based Kriging and BHM, and spatial local heterogeneity based hotspots and geographical regression methods, appraised as the first and second laws of Geography (Tobler 1970; Goodchild 2004), respectively. Spatial stratified heterogeneity (SSH), the phenomena of a partition that within strata is more similar than between strata, examples are climate zones and landuse classes and remote sensing classification, is prevalent in geography and understood since ancient Greek, is surprisingly neglected in Spatial Statistics, probably due to the existence of hundreds of classification algorithms. In this article, we go beyond the classifications and disclose that SSH is the sources of sample bias, statistic bias, modelling confounding and misleading CI, and recommend robust solutions to overcome the negativity. In the meantime, we elaborate four benefits from SSH: creating identical PDF or equivalent to random sampling in stratum; the spatial pattern in strata, the borders between strata as a specific information for nonlinear causation; and general interaction by overlaying two spatial patterns. We developed the equation of SSH and discuss its context. The comprehensive investigation formulates the statistics for SSH, presenting a new principle and toolbox in spatial statistics.

stat.ME↗

arXiv2022

Stepwise Feature Fusion: Local Guides Global

Colonoscopy, currently the most efficient and recognized colon polyp detection technology, is necessary for early screening and prevention of colorectal cancer. However, due to the varying size and complex morphological features of colonic polyps as well as the indistinct boundary between polyps and mucosa, accurate segmentation of polyps is still challenging. Deep learning has become popular for accurate polyp segmentation tasks with excellent results. However, due to the structure of polyps image and the varying shapes of polyps, it easy for existing deep learning models to overfitting the current dataset. As a result, the model may not process unseen colonoscopy data. To address this, we propose a new State-Of-The-Art model for medical image segmentation, the SSFormer, which uses a pyramid Transformer encoder to improve the generalization ability of models. Specifically, our proposed Progressive Locality Decoder can be adapted to the pyramid Transformer backbone to emphasize local features and restrict attention dispersion. The SSFormer achieves statet-of-the-art performance in both learning and generalization assessment.

eess.IV↗

arXiv2021

Seasonal association between viral causes of hospitalised acute lower respiratory infections and meteorological factors in China: a retrospective study

Acute lower respiratory infections caused by respiratory viruses are common and persistent infectious diseases worldwide and in China, which have pronounced seasonal patterns. Meteorological factors have important roles in the seasonality of some major viruses. Our aim was to identify the dominant meteorological factors and to model their effects on common respiratory viruses in different regions of China. We analysed monthly virus data on patients from 81 sentinel hospitals in 22 provinces in mainland China from 2009 to 2013. The geographical detector method was used to quantify the explanatory power of each meteorological factor, individually and interacting in pairs. 28369 hospitalised patients with ALRI were tested, 10387 were positive for at least one virus, including RSV, influenza virus, PIV, ADV, hBoV, hCoV and hMPV. RSV and influenza virus had annual peaks in the north and biannual peaks in the south. PIV and hBoV had higher positive rates in the spring summer months. hMPV had an annual peak in winter spring, especially in the north. ADV and hCoV exhibited no clear annual seasonality. Temperature, atmospheric pressure, vapour pressure, and rainfall had most explanatory power on most respiratory viruses in each region. Relative humidity was only dominant in the north, but had no significant explanatory power for most viruses in the south. Hours of sunlight had significant explanatory power for RSV and influenza virus in the north, and for most viruses in the south. Wind speed was the only factor with significant explanatory power for human coronavirus in the south. For all viruses, interactions between any two of the paired factors resulted in enhanced explanatory power, either bivariately or non-linearly.

stat.AP↗

arXiv2021

Enhancing Population Persistence by a Protection Zone in a Reaction-Diffusion Model with Strong Allee Effect

Protecting endangered species has been an important issue in ecology. We derive a reaction-diffusion model for a population in a one-dimensional bounded habitat, where the population is subjected to a strong Allee effect in its natural domain but obeys a logistic growth in a protection zone. We establish the conditions for population persistence and extinction via the principal eigenvalue of an associated eigenvalue problem and investigate the dependence of this principal eigenvalue on the location (i.e., the starting point and the length) of the protection zone. The results are used to design the optimal protection zone under different boundary conditions, that is, to suggest the starting point and length of the protection zone with respect to different population growth rate in the protection zone, in order for the population to persist in a long term.

math.DS↗