Yihua Zhang
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Research Interests
Yihua Zhang has a total of 13 co-authors. Here are the top co-authors by number of shared publications:
- Sijia Liu
Michigan State University
4 shared publications - Changsheng Wang
Michigan State University
2 shared publications - Pin-Yu Chen
IBM Research, Yorktown Heights, New York, USA
2 shared publications - Dennis Wei
IBM Research, San Jose, USA
1 shared publication Haomin Zhuang
1 shared publication- Hongkang Li
Department of Electrical, Computer, and Systems Engineering, Rensselaer Polytechnic Institute, Troy, NY, USA
1 shared publication
Latest Publications
- 1 May 2026
- Published by Springer Nature
- p. 77-88
Data Integrity for Machine Unlearning
- Open Access
- 13 October 2025
- Published by Association for Computing Machinery (ACM)
- p. 136-145
LLM Unlearning on Noisy Forget Sets: A Study of Incomplete, Rewritten, and Watermarked Data
- 13 October 2025
- Published by Association for Computing Machinery (ACM)
- p. 136-145
Open AccessLLM Unlearning on Noisy Forget Sets: A Study of Incomplete, Rewritten, and Watermarked Data
- 7 March 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal of Selected Topics in Signal Processing
- Vol. 18 (2) , 216-231
How Does Promoting the Minority Fraction Affect Generalization? A Theoretical Study of One-Hidden-Layer Neural Network on Group Imbalance
Group imbalance has been a known problem in empirical risk minimization (ERM), where the achieved high average accuracy is accompanied by low accuracy in a minority group. Despite algorithmic efforts to improve the minority group accuracy, a theoretical generalization analysis of ERM on individual groups remains elusive. By formulating the group imbalance problem with the Gaussian Mixture Model, this paper quantifies the impact of individual groups on the sample complexity, the convergence rate, and the average and group-level testing performance. Although our theoretical framework is centered on binary classification using a one-hidden-layer neural network, to the best of our knowledge, we provide the first theoretical analysis of the group-level generalization of ERM in addition to the commonly studied average generalization performance. Sample insights of our theoretical results include that when all group-level co-variance is in the medium regime and all mean are close to zero, the learning performance is most desirable in the sense of a small sample complexity, a fast training rate, and a high average and group-level testing accuracy. Moreover, we show that increasing the fraction of the minority group in the training data does not necessarily improve the generalization performance of the minority group. Our theoretical results are validated on both synthetic and empirical datasets, such as CelebA and CIFAR-10 in image classification.
- 7 March 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal of Selected Topics in Signal Processing
- Vol. 18 (2) , 216-231
How Does Promoting the Minority Fraction Affect Generalization? A Theoretical Study of One-Hidden-Layer Neural Network on Group Imbalance
Group imbalance has been a known problem in empirical risk minimization (ERM), where the achieved high average accuracy is accompanied by low accuracy in a minority group. Despite algorithmic efforts to improve the minority group accuracy, a theoretical generalization analysis of ERM on individual groups remains elusive. By formulating the group imbalance problem with the Gaussian Mixture Model, this paper quantifies the impact of individual groups on the sample complexity, the convergence rate, and the average and group-level testing performance. Although our theoretical framework is centered on binary classification using a one-hidden-layer neural network, to the best of our knowledge, we provide the first theoretical analysis of the group-level generalization of ERM in addition to the commonly studied average generalization performance. Sample insights of our theoretical results include that when all group-level co-variance is in the medium regime and all mean are close to zero, the learning performance is most desirable in the sense of a small sample complexity, a fast training rate, and a high average and group-level testing accuracy. Moreover, we show that increasing the fraction of the minority group in the training data does not necessarily improve the generalization performance of the minority group. Our theoretical results are validated on both synthetic and empirical datasets, such as CelebA and CIFAR-10 in image classification.
- Top Cited Papers
- 1 January 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Signal Processing Magazine
- Vol. 41 (1) , 38-59
An Introduction to Bilevel Optimization: Foundations and applications in signal processing and machine learning
Recently, bilevel optimization (BLO) has taken center stage in some very exciting developments in the area of signal processing (SP) and machine learning (ML). Roughly speaking, BLO is a classical optimization problem that involves two levels of hierarchy (i.e., upper and lower levels), wherein obtaining the solution to the upper-level problem requires solving the lower-level one. BLO has become popular largely because it is powerful in modeling problems in SP and ML, among others, that involve optimizing nested objective functions. Prominent applications of BLO range from resource allocation for wireless systems to adversarial ML. In this work, we focus on a class of tractable BLO problems that often appear in SP and ML applications. We provide an overview of some basic concepts of this class of BLO problems, such as their optimality conditions, standard algorithms (including their optimization principles and practical implementations) as well as how they can be leveraged to obtain state-of-the-art results for several key SP and ML applications. Further, we discuss some recent advances in BLO theory and its implications for applications, and we point out some limitations of the state of the art that require significant future research efforts. We hope that this article, together with the associated open source BLO toolbox we developed for algorithm benchmarking, can serve to accelerate the adoption of BLO as a generic tool to model, analyze, and innovate on a wide array of emerging SP and ML applications.
- 1 January 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Signal Processing Magazine
- Vol. 41 (1) , 38-59
Top Cited PapersAn Introduction to Bilevel Optimization: Foundations and applications in signal processing and machine learning
Recently, bilevel optimization (BLO) has taken center stage in some very exciting developments in the area of signal processing (SP) and machine learning (ML). Roughly speaking, BLO is a classical optimization problem that involves two levels of hierarchy (i.e., upper and lower levels), wherein obtaining the solution to the upper-level problem requires solving the lower-level one. BLO has become popular largely because it is powerful in modeling problems in SP and ML, among others, that involve optimizing nested objective functions. Prominent applications of BLO range from resource allocation for wireless systems to adversarial ML. In this work, we focus on a class of tractable BLO problems that often appear in SP and ML applications. We provide an overview of some basic concepts of this class of BLO problems, such as their optimality conditions, standard algorithms (including their optimization principles and practical implementations) as well as how they can be leveraged to obtain state-of-the-art results for several key SP and ML applications. Further, we discuss some recent advances in BLO theory and its implications for applications, and we point out some limitations of the state of the art that require significant future research efforts. We hope that this article, together with the associated open source BLO toolbox we developed for algorithm benchmarking, can serve to accelerate the adoption of BLO as a generic tool to model, analyze, and innovate on a wide array of emerging SP and ML applications.
- Open Access
A Pilot Study of Query-Free Adversarial Attack against Stable Diffusion
Despite the record-breaking performance in Text-to-Image (T2I) generation by Stable Diffusion, less research attention is paid to its adversarial robustness. In this work, we study the problem of adversarial attack generation for Stable Diffusion and ask if an adversarial text prompt can be obtained even in the absence of end-to-end model queries. We call the resulting problem 'query-free attack generation'. To resolve this problem, we show that the vulnerability of T2I models is rooted in the lack of robustness of text encoders, e.g., the CLIP text encoder used for attacking Stable Diffusion. Based on such insight, we propose both untargeted and targeted query-free attacks, where the former is built on the most influential dimensions in the text embedding space, which we call steerable key dimensions. By leveraging the proposed attacks, we empirically show that only a five-character perturbation to the text prompt is able to cause the significant content shift of synthesized images using Stable Diffusion. Moreover, we show that the proposed target attack can precisely steer the diffusion model to scrub the targeted image content without causing much change in untargeted image content. Our code is available at https://github.com/OPTML-Group/QF-Attack.
Open AccessA Pilot Study of Query-Free Adversarial Attack against Stable Diffusion
Despite the record-breaking performance in Text-to-Image (T2I) generation by Stable Diffusion, less research attention is paid to its adversarial robustness. In this work, we study the problem of adversarial attack generation for Stable Diffusion and ask if an adversarial text prompt can be obtained even in the absence of end-to-end model queries. We call the resulting problem 'query-free attack generation'. To resolve this problem, we show that the vulnerability of T2I models is rooted in the lack of robustness of text encoders, e.g., the CLIP text encoder used for attacking Stable Diffusion. Based on such insight, we propose both untargeted and targeted query-free attacks, where the former is built on the most influential dimensions in the text embedding space, which we call steerable key dimensions. By leveraging the proposed attacks, we empirically show that only a five-character perturbation to the text prompt is able to cause the significant content shift of synthesized images using Stable Diffusion. Moreover, we show that the proposed target attack can precisely steer the diffusion model to scrub the targeted image content without causing much change in untargeted image content. Our code is available at https://github.com/OPTML-Group/QF-Attack.