Lingyan Bian
Research Subjects
Lingyan Bian has a total of 7 co-authors. Here are the top co-authors by number of shared publications:
- Ang Li
School of Petroleum, China University of Petroleum (Beijing) at Karamay Campus, Karamay 834000, China
1 shared publication - Chengyin Hu
China University of Petroleum‐Beijing at Karamay Karamay Xinjiang China
1 shared publication - Haitao Shi
School of Mathematics, Physics and Artificial Intelligence, China University of Petroleum–Beijing at Karamay, Karamay 834000, China
1 shared publication - Tianle Liang
China University of Petroleum Beijing at Karamay, Karamay, China
1 shared publication - Xuelian Shi
China University of Petroleum Beijing at Karamay, Karamay, China
1 shared publication - Yijun Chen
China University of Petroleum Beijing at Karamay, Karamay, China
1 shared publication
Latest Publications
- Open Access
- 15 June 2026
- Published by Association for Computing Machinery (ACM)
- p. 1974-1983
GMAE: Gated Multi-signal Adaptive Ensemble for Targeted Transfer Attacks on Commercial Large Vision-Language Models
Despite strong performance on open-source large vision-language models (LVLMs), transfer-based targeted adversarial attacks often fail against closed-source commercial LVLMs. Existing methods typically rely on static perturbation allocation and uniformly weighted surrogate ensembles, limiting their ability to capture model-agnostic semantic vulnerabilities essential for cross-model transferability. Consequently, perturbations are frequently allocated to semantically irrelevant regions, reducing effectiveness under realistic black-box settings. To address this limitation, we propose the Gated Multi-signal Adaptive Ensemble (GMAE) framework, a novel transfer-based targeted attack that formulates adversarial generation as a joint optimization over a gated perturbation field and an adaptive ensemble objective. A learnable spatial gating mask dynamically reallocates perturbation budgets toward semantically vulnerable regions, while surrogate models are adaptively reweighted according to optimization difficulty and historical consistency, encouraging alignment on shared semantic weaknesses across heterogeneous architectures. To further enhance robustness, GMAE performs global semantic alignment in a symbolic discrete space, reducing the influence of non-essential feature variations and improving cross-model generalization. In addition, we introduce a local attention-guided multi-scale consistency mechanism that transfers fine-grained target textures into attention-identified sensitive regions, ensuring both visual imperceptibility and semantic coherence at feature and concept levels. Extensive experiments demonstrate that GMAE achieves strong transferability to commercial LVLMs, and reasoning-oriented models. Notably, GMAE achieves an attack success rate of 82.90% on GPT-4o, outperforming the strongest baseline by 7.8%, establishing a new state of the art for targeted transfer attacks on large-scale black-box LVLMs.
- 15 June 2026
- Published by Association for Computing Machinery (ACM)
- p. 1974-1983
Open AccessGMAE: Gated Multi-signal Adaptive Ensemble for Targeted Transfer Attacks on Commercial Large Vision-Language Models
Despite strong performance on open-source large vision-language models (LVLMs), transfer-based targeted adversarial attacks often fail against closed-source commercial LVLMs. Existing methods typically rely on static perturbation allocation and uniformly weighted surrogate ensembles, limiting their ability to capture model-agnostic semantic vulnerabilities essential for cross-model transferability. Consequently, perturbations are frequently allocated to semantically irrelevant regions, reducing effectiveness under realistic black-box settings. To address this limitation, we propose the Gated Multi-signal Adaptive Ensemble (GMAE) framework, a novel transfer-based targeted attack that formulates adversarial generation as a joint optimization over a gated perturbation field and an adaptive ensemble objective. A learnable spatial gating mask dynamically reallocates perturbation budgets toward semantically vulnerable regions, while surrogate models are adaptively reweighted according to optimization difficulty and historical consistency, encouraging alignment on shared semantic weaknesses across heterogeneous architectures. To further enhance robustness, GMAE performs global semantic alignment in a symbolic discrete space, reducing the influence of non-essential feature variations and improving cross-model generalization. In addition, we introduce a local attention-guided multi-scale consistency mechanism that transfers fine-grained target textures into attention-identified sensitive regions, ensuring both visual imperceptibility and semantic coherence at feature and concept levels. Extensive experiments demonstrate that GMAE achieves strong transferability to commercial LVLMs, and reasoning-oriented models. Notably, GMAE achieves an attack success rate of 82.90% on GPT-4o, outperforming the strongest baseline by 7.8%, establishing a new state of the art for targeted transfer attacks on large-scale black-box LVLMs.