Chuan Ma

AAAI Conference 2026 Conference Paper

TVChain: Leveraging Textual-Visual Prompt Chains for Jailbreaking Large Vision-Language Models

• Hao Yu
• Ke Liang
• Junxian Duan
• Jun Wang
• Siwei Wang
• Chuan Ma
• Xinwang Liu

Large Vision-Language Models (LVLMs) enhance the capabilities of Large Language Models by integrating visual inputs, thereby enabling advanced multimodal reasoning across diverse applications. However, these enhanced reasoning capabilities introduce new security risks, particularly to jailbreaking attacks that bypass built-in safety mechanisms to elicit harmful or unauthorized outputs. While recent efforts have explored adversarial and typographic prompts, most existing attacks suffer from three key limitations: reliance on auxiliary models, limited effectiveness in black-box scenarios, and inadequate exploitation of the LVLMs' intrinsic reasoning abilities. In this work, we propose TVChain, a novel black-box jailbreaking framework that explicitly intervenes in both the visual and textual reasoning processes of LVLMs. TVChain decomposes malicious prompts into a sequence of semantically meaningful sub-images that represent relevant objects and behaviors, thereby circumventing direct exposure of illicit content. In parallel, a carefully designed chain-of-thought (CoT) textual prompt is employed to steer the model's reasoning toward reconstructing the intended activity in a covert yet effective manner. We demonstrate that this compositional prompting strategy reduces the likelihood of triggering safety mechanisms while preserving attack efficacy. Extensive evaluations on eleven LVLMs (seven open-source and four commercial) across two benchmark datasets and three state-of-the-art defenses validate the effectiveness and robustness of TVChain.

IJCAI Conference 2025 Conference Paper

Advancing Embodied Agent Security: From Safety Benchmarks to Input Moderation

• Ning Wang
• Zihan Yan
• Weiyang Li
• Chuan Ma
• He Chen
• Tao Xiang

Embodied agents exhibit immense potential across a multitude of domains, making the assurance of their behavioral safety a fundamental prerequisite for their widespread deployment. However, existing research predominantly concentrates on the security of general large language models, lacking specialized methodologies for establishing safety benchmarks and input moderation tailored to embodied agents. To bridge this gap, this paper introduces a novel input moderation framework, meticulously designed to safeguard embodied agents. This framework encompasses the entire pipeline, including taxonomy definition, dataset curation, moderator architecture, model training, and rigorous evaluation. Notably, we introduce EAsafetyBench, a meticulously crafted safety benchmark engineered to facilitate both the training and stringent assessment of moderators specifically designed for embodied agents. Furthermore, we propose Pinpoint, an innovative prompt-decoupled input moderation scheme that harnesses a masked attention mechanism to effectively isolate and mitigate the influence of functional prompts on moderation tasks. Extensive experiments conducted on diverse benchmark datasets and models validate the feasibility and efficacy of the proposed approach. The results demonstrate that our methodologies achieve an impressive average detection accuracy of 94. 58%, surpassing the performance of existing state-of-the-art techniques, alongside an exceptional moderation processing time of merely 0. 002 seconds per instance. The source code and datasets can be found at https: //github. com/ZihanYan-CQU/EAsafetyBench.

EAAI Journal 2025 Journal Article

An improved framework for breast ultrasound image segmentation with multiple branches depth perception and layer compression residual module

• Ke Cui
• Qichuan Tian
• Haoji Wang
• Chuan Ma

Breast cancer is becoming a leading cause of death among women worldwide. Early detection is essential for improving survival rates and facilitating targeted medical treatments. Automated segmentation of breast tumors from ultrasound images is vital for this early diagnosis. To tackle challenges such as low contrast, unclear lesion boundaries, and class imbalance in breast ultrasound images, a multiple branches depth perception network was introduced, using a symmetric encoder and decoder architecture. In the initial feature extraction stage, the network’s encoder employs the multiple branches depth residual block to integrate feature information from multiple branches while employing dilated convolution to capture intricate contextual details, enhancing the characterization of complex features. Subsequently, in the feature recovery stage, the network utilizes dual path depth perception block to mitigate information loss in deep networks by leveraging dual path residual connections, extracting rich textural and structural features from breast ultrasound images. Furthermore, the layer compression residual module and attention refinement module were incorporated within the skip connections to strengthen the contextual relationships between the encoder and decoder, leading to improved segmentation of breast lesions. Extensive qualitative and quantitative evaluations on two challenging public datasets were conducted to assess the effectiveness and generalizability of the proposed approach. The experimental results demonstrate the reliability of the proposed method in clinical treatment, achieving segmentation mean intersection over union scores of 91. 11% and 92. 28% on these respective datasets.

AAAI Conference 2024 Conference Paper

Attribute-Missing Graph Clustering Network

• Wenxuan Tu
• Renxiang Guan
• Sihang Zhou
• Chuan Ma
• Xin Peng
• Zhiping Cai
• Zhe Liu
• Jieren Cheng

Deep clustering with attribute-missing graphs, where only a subset of nodes possesses complete attributes while those of others are missing, is an important yet challenging topic in various practical applications. It has become a prevalent learning paradigm in existing studies to perform data imputation first and subsequently conduct clustering using the imputed information. However, these ``two-stage" methods disconnect the clustering and imputation processes, preventing the model from effectively learning clustering-friendly graph embedding. Furthermore, they are not tailored for clustering tasks, leading to inferior clustering results. To solve these issues, we propose a novel Attribute-Missing Graph Clustering (AMGC) method to alternately promote clustering and imputation in a unified framework, where we iteratively produce the clustering-enhanced nearest neighbor information to conduct the data imputation process and utilize the imputed information to implicitly refine the clustering distribution through model optimization. Specifically, in the imputation step, we take the learned clustering information as imputation prompts to help each attribute-missing sample gather highly correlated features within its clusters for data completion, such that the intra-class compactness can be improved. Moreover, to support reliable clustering, we maximize inter-class separability by conducting cost-efficient dual non-contrastive learning over the imputed latent features, which in turn promotes greater graph encoding capability for clustering sub-network. Extensive experiments on five datasets have verified the superiority of AMGC against competitors.

AAMAS Conference 2004 Conference Paper

Multiagent Planning as Control Synthesis

• Kiam Tian Seow
• Chuan Ma
• Makoto Yokoo