Yunqi Zhang
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Yunqi Zhang has a total of 26 co-authors. Here are the top co-authors by number of shared publications:
- Liwen Wu
Engineering Research Center of Cyberspace, Yunnan University, Kunming 650500, China
2 shared publications Feng Wu
Engineering Research Center of Cyberspace, Yunnan University, Kunming, China
1 shared publication- Ji-Bin Li
State Key Laboratory of Oncology in South China, Collaborative Innovation Center for Cancer Medicine, Guangzhou, China
1 shared publication - Jiayu Li
College of Electrical Engineering and New Energy, China Three Gorges University, Yichang, China
1 shared publication Jie Li
Engineering Research Center of Cyberspace, Yunnan University, Kunming, China
1 shared publication- Jieyi Yi
School of Urban Governance and Public Affairs, Suzhou City University, Suzhou 215104, China
1 shared publication
Latest Publications
- Open Access
Structured Spike-and-Slab Variational Bayes for High-Dimensional Non-Normal Generalized Linear Mixed Models
High-dimensional non-normal longitudinal data are ubiquitous across fields such as genomics, biomedicine, microbiome research, and the social sciences. Such data often combine non-normal responses, within-subject dependence and sparse population effects. We develop a structured variational Bayesian procedure: variational Bayesian empirical likelihood with spike-and-slab priors, which combines an empirical-likelihood-motivated working kernel with two Laplace shrinkage branches. The variational family links each inclusion indicator to its coefficient and local scale, and it uses probability-weighted updates for global shrinkage. A scalar empirical-likelihood weighting implementation provides tractable computation, and conditional working-evidence scores compare random-effect covariance structures. Simulation studies examine selection, estimation, prediction, interval coverage, and sensitivity to numerical and prior settings. Two longitudinal microbiome applications yield interpretable conditional associations. The resulting framework provides an explicit computational construction for sparse mixed-model analysis, with numerical results characterizing its operating behavior.
Open AccessStructured Spike-and-Slab Variational Bayes for High-Dimensional Non-Normal Generalized Linear Mixed Models
High-dimensional non-normal longitudinal data are ubiquitous across fields such as genomics, biomedicine, microbiome research, and the social sciences. Such data often combine non-normal responses, within-subject dependence and sparse population effects. We develop a structured variational Bayesian procedure: variational Bayesian empirical likelihood with spike-and-slab priors, which combines an empirical-likelihood-motivated working kernel with two Laplace shrinkage branches. The variational family links each inclusion indicator to its coefficient and local scale, and it uses probability-weighted updates for global shrinkage. A scalar empirical-likelihood weighting implementation provides tractable computation, and conditional working-evidence scores compare random-effect covariance structures. Simulation studies examine selection, estimation, prediction, interval coverage, and sensitivity to numerical and prior settings. Two longitudinal microbiome applications yield interpretable conditional associations. The resulting framework provides an explicit computational construction for sparse mixed-model analysis, with numerical results characterizing its operating behavior.
- Open Access
- 4 July 2026
- Published by MDPI AG in Mathematics
- Vol. 14 (13) , 2392
Boosting the Transferability of Adversarial Attacks with Semantic-Invariance and Low-Gradient Replacement
Deepneural networks are highly vulnerable to adversarial examples, which are generated by introducing subtle perturbations to input data to mislead model classification. Currently, transfer-based attacks are prevalent in adversarial example generation and can be categorized into input transformation-based and gradient-based methods. However, most input transformation-based methods tend to produce augmented replicas that are semantically inconsistent with the original inputs, while gradient-based methods often leave low-gradient regions unperturbed within the model’s critical attention areas. These limitations constrain further improvements in adversarial transferability. In this work, we propose a Semantic-Invariance and Low-Gradient Replacement Method (SLRM) to address these challenges. Our framework integrates semantically consistent augmentation and gradient replacement as follows: (1) a feature extractor captures semantic features from original inputs and a reconstructor generates augmented replicas that preserve semantic fidelity to enhance input diversity, and (2) low-gradient regions in adversarial examples are systematically replaced with corresponding regions from augmented replicas to eliminate the under-perturbed areas critical for model robustness. Comprehensive empirical evaluations on ImageNet demonstrate that SLRM significantly enhances the transferability of baseline methods and seamlessly integrates with state-of-the-art approaches to further improve their performance. Moreover, SLRM substantially improves the robustness of baseline methods against defended models, achieving superior attack success rates under advanced adversarial defenses.
- 4 July 2026
- Published by MDPI AG in Mathematics
- Vol. 14 (13) , 2392
Open AccessBoosting the Transferability of Adversarial Attacks with Semantic-Invariance and Low-Gradient Replacement
Deepneural networks are highly vulnerable to adversarial examples, which are generated by introducing subtle perturbations to input data to mislead model classification. Currently, transfer-based attacks are prevalent in adversarial example generation and can be categorized into input transformation-based and gradient-based methods. However, most input transformation-based methods tend to produce augmented replicas that are semantically inconsistent with the original inputs, while gradient-based methods often leave low-gradient regions unperturbed within the model’s critical attention areas. These limitations constrain further improvements in adversarial transferability. In this work, we propose a Semantic-Invariance and Low-Gradient Replacement Method (SLRM) to address these challenges. Our framework integrates semantically consistent augmentation and gradient replacement as follows: (1) a feature extractor captures semantic features from original inputs and a reconstructor generates augmented replicas that preserve semantic fidelity to enhance input diversity, and (2) low-gradient regions in adversarial examples are systematically replaced with corresponding regions from augmented replicas to eliminate the under-perturbed areas critical for model robustness. Comprehensive empirical evaluations on ImageNet demonstrate that SLRM significantly enhances the transferability of baseline methods and seamlessly integrates with state-of-the-art approaches to further improve their performance. Moreover, SLRM substantially improves the robustness of baseline methods against defended models, achieving superior attack success rates under advanced adversarial defenses.
- 25 January 2025
- Published by Elsevier in Journal of Information Security and Applications
- Vol. 89
CSA: Crafting adversarial examples via content and style attacks
- 25 January 2025
- Published by Elsevier in Journal of Information Security and Applications
- Vol. 89
CSA: Crafting adversarial examples via content and style attacks
- 17 July 2024
- Published by Springer Nature in Communications in Mathematics and Statistics
- Vol. 14 (4) , 1-45
Equivalence Assessment via the Difference Between Two AUCs in a Matched-Pair Design with Nonignorable Missing Endpoints
- 17 July 2024
- Published by Springer Nature in Communications in Mathematics and Statistics
- Vol. 14 (4) , 1-45
Equivalence Assessment via the Difference Between Two AUCs in a Matched-Pair Design with Nonignorable Missing Endpoints
- Open Access
- 3 January 2023
- Published by MDPI AG in Mathematics
- Vol. 11 (1) , 246
Non-Parametric Non-Inferiority Assessment in a Three-Arm Trial with Non-Ignorable Missing Data
A three-arm non-inferiority trial including a placebo is usually utilized to assess the non-inferiority of an experimental treatment to a reference treatment. Existing methods for assessing non-inferiority mainly focus on the fully observed endpoints. However, in some clinical trials, treatment endpoints may be subject to missingness for various reasons, such as the refusal of subjects or their migration. To address this issue, this paper aims to develop a non-parametric approach to assess the non-inferiority of an experimental treatment to a reference treatment in a three-arm trial with non-ignorable missing endpoints. A logistic regression is adopted to specify a non-ignorable missingness data mechanism. A semi-parametric imputation method is proposed to estimate parameters in the considered logistic regression. Inverse probability weighting, augmented inverse probability weighting and non-parametric methods are developed to estimate treatment efficacy for known and unknown parameters in the considered logistic regression. Under some regularity conditions, we show asymptotic normality of the constructed estimators for treatment efficacy. A bootstrap resampling method is presented to estimate asymptotic variances of the estimated treatment efficacy. Three Wald-type statistics are constructed to test the non-inferiority based on the asymptotic properties of the estimated treatment efficacy. Empirical studies show that the proposed Wald-type test procedure is robust to the misspecified missingness data mechanism, and behaves better than the complete-case method in the sense that the type I error rates for the former are closer to the pre-given significance level than those for the latter.
- 3 January 2023
- Published by MDPI AG in Mathematics
- Vol. 11 (1) , 246
Open AccessNon-Parametric Non-Inferiority Assessment in a Three-Arm Trial with Non-Ignorable Missing Data
A three-arm non-inferiority trial including a placebo is usually utilized to assess the non-inferiority of an experimental treatment to a reference treatment. Existing methods for assessing non-inferiority mainly focus on the fully observed endpoints. However, in some clinical trials, treatment endpoints may be subject to missingness for various reasons, such as the refusal of subjects or their migration. To address this issue, this paper aims to develop a non-parametric approach to assess the non-inferiority of an experimental treatment to a reference treatment in a three-arm trial with non-ignorable missing endpoints. A logistic regression is adopted to specify a non-ignorable missingness data mechanism. A semi-parametric imputation method is proposed to estimate parameters in the considered logistic regression. Inverse probability weighting, augmented inverse probability weighting and non-parametric methods are developed to estimate treatment efficacy for known and unknown parameters in the considered logistic regression. Under some regularity conditions, we show asymptotic normality of the constructed estimators for treatment efficacy. A bootstrap resampling method is presented to estimate asymptotic variances of the estimated treatment efficacy. Three Wald-type statistics are constructed to test the non-inferiority based on the asymptotic properties of the estimated treatment efficacy. Empirical studies show that the proposed Wald-type test procedure is robust to the misspecified missingness data mechanism, and behaves better than the complete-case method in the sense that the type I error rates for the former are closer to the pre-given significance level than those for the latter.
- Open Access
- 30 November 2022
- Published by MDPI AG in Mathematics
- Vol. 10 (23) , 4525
Hateful Memes Detection Based on Multi-Task Learning
With the popularity of posting memes on social platforms, the severe negative impact of hateful memes is growing. As existing detection models have lower detection accuracy than humans, hateful memes detection is still a challenge to statistical learning and artificial intelligence. This paper proposed a multi-task learning method consisting of a primary multimodal task and two unimodal auxiliary tasks to address this issue. We introduced a self-supervised generation strategy in auxiliary tasks to generate unimodal auxiliary labels automatically. Meanwhile, we used BERT and RESNET as the backbone for text and image classification, respectively, and then fusion them with a late fusion method. In the training phase, the backward guidance technique and the adaptive weight adjustment strategy were used to capture the consistency and variability between different modalities, numerically improving the hateful memes detection accuracy and the generalization and robustness of the model. The experiment conducted on the Facebook AI multimodal hateful memes dataset shows that the prediction accuracy of our model outperformed the comparing models.
- 30 November 2022
- Published by MDPI AG in Mathematics
- Vol. 10 (23) , 4525
Open AccessHateful Memes Detection Based on Multi-Task Learning
With the popularity of posting memes on social platforms, the severe negative impact of hateful memes is growing. As existing detection models have lower detection accuracy than humans, hateful memes detection is still a challenge to statistical learning and artificial intelligence. This paper proposed a multi-task learning method consisting of a primary multimodal task and two unimodal auxiliary tasks to address this issue. We introduced a self-supervised generation strategy in auxiliary tasks to generate unimodal auxiliary labels automatically. Meanwhile, we used BERT and RESNET as the backbone for text and image classification, respectively, and then fusion them with a late fusion method. In the training phase, the backward guidance technique and the adaptive weight adjustment strategy were used to capture the consistency and variability between different modalities, numerically improving the hateful memes detection accuracy and the generalization and robustness of the model. The experiment conducted on the Facebook AI multimodal hateful memes dataset shows that the prediction accuracy of our model outperformed the comparing models.
- Open Access
- 15 June 2022
- Published by Frontiers Media SA in Frontiers in Genetics
- Vol. 13, 912614
Gm-PLoc: A Subcellular Localization Model of Multi-Label Protein Based on GAN and DeepFM
Identifying the subcellular localization of a given protein is an essential part of biological and medical research, since the protein must be localized in the correct organelle to ensure physiological function. Conventional biological experiments for protein subcellular localization have some limitations, such as high cost and low efficiency, thus massive computational methods are proposed to solve these problems. However, some of these methods need to be improved further for protein subcellular localization with class imbalance problem. We propose a new model, generating minority samples for protein subcellular localization (Gm-PLoc), to predict the subcellular localization of multi-label proteins. This model includes three steps: using the position specific scoring matrix to extract distinguishable features of proteins; synthesizing samples of the minority category to balance the distribution of categories based on the revised generative adversarial networks; training a classifier with the rebalanced dataset to predict the subcellular localization of multi-label proteins. One benchmark dataset is selected to evaluate the performance of the presented model, and the experimental results demonstrate that Gm-PLoc performs well for the multi-label protein subcellular localization.
- 15 June 2022
- Published by Frontiers Media SA in Frontiers in Genetics
- Vol. 13, 912614
Open AccessGm-PLoc: A Subcellular Localization Model of Multi-Label Protein Based on GAN and DeepFM
Identifying the subcellular localization of a given protein is an essential part of biological and medical research, since the protein must be localized in the correct organelle to ensure physiological function. Conventional biological experiments for protein subcellular localization have some limitations, such as high cost and low efficiency, thus massive computational methods are proposed to solve these problems. However, some of these methods need to be improved further for protein subcellular localization with class imbalance problem. We propose a new model, generating minority samples for protein subcellular localization (Gm-PLoc), to predict the subcellular localization of multi-label proteins. This model includes three steps: using the position specific scoring matrix to extract distinguishable features of proteins; synthesizing samples of the minority category to balance the distribution of categories based on the revised generative adversarial networks; training a classifier with the rebalanced dataset to predict the subcellular localization of multi-label proteins. One benchmark dataset is selected to evaluate the performance of the presented model, and the experimental results demonstrate that Gm-PLoc performs well for the multi-label protein subcellular localization.
- Open Access
- 1 May 2021
- Published by Elsevier in Molecular and Cellular Endocrinology
- Vol. 527, 111218
Maternal obesity interrupts the coordination of the unfolded protein response and heat shock response in the postnatal developing hypothalamus of male offspring in mice
- 1 May 2021
- Published by Elsevier in Molecular and Cellular Endocrinology
- Vol. 527, 111218
Open AccessMaternal obesity interrupts the coordination of the unfolded protein response and heat shock response in the postnatal developing hypothalamus of male offspring in mice