arXiv subjects

Cen Chen

Publications and source records attributed to Cen Chen.

At least 55 records · Page 3Linked to original sources

arXiv2025

Refiner: Data Refining against Gradient Leakage Attacks in Federated Learning

Recent works have brought attention to the vulnerability of Federated Learning (FL) systems to gradient leakage attacks. Such attacks exploit clients' uploaded gradients to reconstruct their sensitive data, thereby compromising the privacy protection capability of FL. In response, various defense mechanisms have been proposed to mitigate this threat by manipulating the uploaded gradients. Unfortunately, empirical evaluations have demonstrated limited resilience of these defenses against sophisticated attacks, indicating an urgent need for more effective defenses. In this paper, we explore a novel defensive paradigm that departs from conventional gradient perturbation approaches and instead focuses on the construction of robust data. Intuitively, if robust data exhibits low semantic similarity with clients' raw data, the gradients associated with robust data can effectively obfuscate attackers. To this end, we design Refiner that jointly optimizes two metrics for privacy protection and performance maintenance. The utility metric is designed to promote consistency between the gradients of key parameters associated with robust data and those derived from clients' data, thus maintaining model performance. Furthermore, the privacy metric guides the generation of robust data towards enlarging the semantic gap with clients' data. Theoretical analysis supports the effectiveness of Refiner, and empirical evaluations on multiple benchmark datasets demonstrate the superior defense effectiveness of Refiner at defending against state-of-the-art attacks.

cs.LG↗

arXiv2025

Boosting Gradient Leakage Attacks: Data Reconstruction in Realistic FL Settings

Federated learning (FL) enables collaborative model training among multiple clients without the need to expose raw data. Its ability to safeguard privacy, at the heart of FL, has recently been a hot-button debate topic. To elaborate, several studies have introduced a type of attacks known as gradient leakage attacks (GLAs), which exploit the gradients shared during training to reconstruct clients' raw data. On the flip side, some literature, however, contends no substantial privacy risk in practical FL environments due to the effectiveness of such GLAs being limited to overly relaxed conditions, such as small batch sizes and knowledge of clients' data distributions. This paper bridges this critical gap by empirically demonstrating that clients' data can still be effectively reconstructed, even within realistic FL environments. Upon revisiting GLAs, we recognize that their performance failures stem from their inability to handle the gradient matching problem. To alleviate the performance bottlenecks identified above, we develop FedLeak, which introduces two novel techniques, partial gradient matching and gradient regularization. Moreover, to evaluate the performance of FedLeak in real-world FL environments, we formulate a practical evaluation protocol grounded in a thorough review of extensive FL literature and industry practices. Under this protocol, FedLeak can still achieve high-fidelity data reconstruction, thereby underscoring the significant vulnerability in FL systems and the urgent need for more effective defense methods.

cs.LG↗

arXiv2025

Decompose, Plan in Parallel, and Merge: A Novel Paradigm for Large Language Models based Planning with Multiple Constraints

Despite significant advances in Large Language Models (LLMs), planning tasks still present challenges for LLM-based agents. Existing planning methods face two key limitations: heavy constraints and cascading errors. To address these limitations, we propose a novel parallel planning paradigm, which Decomposes, Plans for subtasks in Parallel, and Merges subplans into a final plan (DPPM). Specifically, DPPM decomposes the complex task based on constraints into subtasks, generates the subplan for each subtask in parallel, and merges them into a global plan. In addition, our approach incorporates a verification and refinement module, enabling error correction and conflict resolution. Experimental results demonstrate that DPPM significantly outperforms existing methods in travel planning tasks.

cs.CL↗

arXiv2025

SEA: Low-Resource Safety Alignment for Multimodal Large Language Models via Synthetic Embeddings

Multimodal Large Language Models (MLLMs) have serious security vulnerabilities.While safety alignment using multimodal datasets consisting of text and data of additional modalities can effectively enhance MLLM's security, it is costly to construct these datasets. Existing low-resource security alignment methods, including textual alignment, have been found to struggle with the security risks posed by additional modalities. To address this, we propose Synthetic Embedding augmented safety Alignment (SEA), which optimizes embeddings of additional modality through gradient updates to expand textual datasets. This enables multimodal safety alignment training even when only textual data is available. Extensive experiments on image, video, and audio-based MLLMs demonstrate that SEA can synthesize a high-quality embedding on a single RTX3090 GPU within 24 seconds. SEA significantly improves the security of MLLMs when faced with threats from additional modalities. To assess the security risks introduced by video and audio, we also introduced a new benchmark called VA-SafetyBench. High attack success rates across multiple MLLMs validate its challenge. Our code and data will be available at https://github.com/ZeroNLP/SEA.

cs.CL↗

arXiv2025

L3A: Label-Augmented Analytic Adaptation for Multi-Label Class Incremental Learning

Class-incremental learning (CIL) enables models to learn new classes continually without forgetting previously acquired knowledge. Multi-label CIL (MLCIL) extends CIL to a real-world scenario where each sample may belong to multiple classes, introducing several challenges: label absence, which leads to incomplete historical information due to missing labels, and class imbalance, which results in the model bias toward majority classes. To address these challenges, we propose Label-Augmented Analytic Adaptation (L3A), an exemplar-free approach without storing past samples. L3A integrates two key modules. The pseudo-label (PL) module implements label augmentation by generating pseudo-labels for current phase samples, addressing the label absence problem. The weighted analytic classifier (WAC) derives a closed-form solution for neural networks. It introduces sample-specific weights to adaptively balance the class contribution and mitigate class imbalance. Experiments on MS-COCO and PASCAL VOC datasets demonstrate that L3A outperforms existing methods in MLCIL tasks. Our code is available at https://github.com/scut-zx/L3A.

cs.CV↗

arXiv2025

PrivacyRestore: Privacy-Preserving Inference in Large Language Models via Privacy Removal and Restoration

The widespread usage of online Large Language Models (LLMs) inference services has raised significant privacy concerns about the potential exposure of private information in user inputs to malicious eavesdroppers. Existing privacy protection methods for LLMs suffer from either insufficient privacy protection, performance degradation, or large inference time overhead. To address these limitations, we propose PrivacyRestore, a plug-and-play method to protect the privacy of user inputs during LLM inference. The server first trains restoration vectors for each privacy span and then release to clients. Privacy span is defined as a contiguous sequence of tokens within a text that contain private information. The client then aggregate restoration vectors of all privacy spans in the input into a single meta restoration vector which is later sent to the server side along with the input without privacy spans.The private information is restored via activation steering during inference. Furthermore, we prove that PrivacyRestore inherently prevents the linear growth of the privacy budget.We create three datasets, covering medical and legal domains, to evaluate the effectiveness of privacy preserving methods. The experimental results show that PrivacyRestore effectively protects private information and maintain acceptable levels of performance and inference overhead.

cs.CR↗

arXiv2025

Reinforced MLLM: A Survey on RL-Based Reasoning in Multimodal Large Language Models

The application of reinforcement learning (RL) to enhance the reasoning capabilities of Multimodal Large Language Models (MLLMs) constitutes a rapidly advancing research area. While MLLMs extend Large Language Models (LLMs) to handle diverse modalities such as vision, audio, and video, enabling robust reasoning across multimodal inputs remains challenging. This paper provides a systematic review of recent advances in RL-based reasoning for MLLMs, covering key algorithmic designs, reward mechanism innovations, and practical applications. We highlight two main RL paradigms, value-model-free and value-model-based methods, and analyze how RL enhances reasoning abilities by optimizing reasoning trajectories and aligning multimodal information. Additionally, we provide an extensive overview of benchmark datasets, evaluation protocols, and current limitations, and propose future research directions to address challenges such as sparse rewards, inefficient cross-modal reasoning, and real-world deployment constraints. Our goal is to provide a comprehensive and structured guide to RL-based multimodal reasoning.

cs.AI↗

arXiv2025

Responsible Diffusion Models via Constraining Text Embeddings within Safe Regions

The remarkable ability of diffusion models to generate high-fidelity images has led to their widespread adoption. However, concerns have also arisen regarding their potential to produce Not Safe for Work (NSFW) content and exhibit social biases, hindering their practical use in real-world applications. In response to this challenge, prior work has focused on employing security filters to identify and exclude toxic text, or alternatively, fine-tuning pre-trained diffusion models to erase sensitive concepts. Unfortunately, existing methods struggle to achieve satisfactory performance in the sense that they can have a significant impact on the normal model output while still failing to prevent the generation of harmful content in some cases. In this paper, we propose a novel self-discovery approach to identifying a semantic direction vector in the embedding space to restrict text embedding within a safe region. Our method circumvents the need for correcting individual words within the input text and steers the entire text prompt towards a safe region in the embedding space, thereby enhancing model robustness against all possibly unsafe prompts. In addition, we employ Low-Rank Adaptation (LoRA) for semantic direction vector initialization to reduce the impact on the model performance for other semantics. Furthermore, our method can also be integrated with existing methods to improve their social responsibility. Extensive experiments on benchmark datasets demonstrate that our method can effectively reduce NSFW content and mitigate social bias generated by diffusion models compared to several state-of-the-art baselines.

cs.CL↗

arXiv2025

Semantic Shift Estimation via Dual-Projection and Classifier Reconstruction for Exemplar-Free Class-Incremental Learning

Exemplar-Free Class-Incremental Learning (EFCIL) aims to sequentially learn from distinct categories without retaining exemplars but easily suffers from catastrophic forgetting of learned knowledge. While existing EFCIL methods leverage knowledge distillation to alleviate forgetting, they still face two critical challenges: semantic shift and decision bias. Specifically, the embeddings of old tasks shift in the embedding space after learning new tasks, and the classifier becomes biased towards new tasks due to training solely with new data, hindering the balance between old and new knowledge. To address these issues, we propose the Dual-Projection Shift Estimation and Classifier Reconstruction (DPCR) approach for EFCIL. DPCR effectively estimates semantic shift through a dual-projection, which combines a learnable transformation with a row-space projection to capture both task-wise and category-wise shifts. Furthermore, to mitigate decision bias, DPCR employs ridge regression to reformulate a classifier reconstruction process. This reconstruction exploits previous in covariance and prototype of each class after calibration with estimated shift, thereby reducing decision bias. Extensive experiments demonstrate that, on various datasets, DPCR effectively balances old and new tasks, outperforming state-of-the-art EFCIL methods. Our codes are available at https://github.com/RHe502/ICML25-DPCR.

cs.CV↗

arXiv2025

Valley: Video Assistant with Large Language model Enhanced abilitY

Large Language Models (LLMs), with remarkable conversational capability, have emerged as AI assistants that can handle both visual and textual modalities. However, their effectiveness in joint video and language understanding has not been extensively explored. In the paper, we introduce Valley, a multi-modal foundation model that is designed to enable enhanced video comprehension and instruction-following capabilities. To this end, we construct two datasets, namely Valley-702k and Valley-instruct-73k, to cover a diverse range of video-text alignment and video-based instruction tasks, such as multi-shot captions, long video descriptions, action recognition, causal inference, etc. Then, we adopt ViT-L/14 as the vision encoder and explore three different temporal modeling modules to learn multifaceted features for enhanced video understanding. In addition, we implement a two-phase training approach for Valley: the first phase focuses solely on training the projection module to facilitate the LLM's capacity to understand visual input, and the second phase jointly trains the projection module and the LLM to improve their instruction following ability. Extensive experiments demonstrate that Valley has the potential to serve as an effective video assistant, simplifying complex video-understanding scenarios. Our code and data are published anonymously at https://github.com/valley-vl/Valley.

cs.CV↗

arXiv2025

Growth Inhibitors for Suppressing Inappropriate Image Concepts in Diffusion Models

Despite their remarkable image generation capabilities, text-to-image diffusion models inadvertently learn inappropriate concepts from vast and unfiltered training data, which leads to various ethical and business risks. Specifically, model-generated images may exhibit not safe for work (NSFW) content and style copyright infringements. The prompts that result in these problems often do not include explicit unsafe words; instead, they contain obscure and associative terms, which are referred to as implicit unsafe prompts. Existing approaches directly fine-tune models under textual guidance to alter the cognition of the diffusion model, thereby erasing inappropriate concepts. This not only requires concept-specific fine-tuning but may also incur catastrophic forgetting. To address these issues, we explore the representation of inappropriate concepts in the image space and guide them towards more suitable ones by injecting growth inhibitors, which are tailored based on the identified features related to inappropriate concepts during the diffusion process. Additionally, due to the varying degrees and scopes of inappropriate concepts, we train an adapter to infer the corresponding suppression scale during the injection process. Our method effectively captures the manifestation of subtle words at the image level, enabling direct and efficient erasure of target concepts without the need for fine-tuning. Through extensive experimentation, we demonstrate that our approach achieves superior erasure results with little effect on other concepts while preserving image quality and semantics.

cs.CV↗

arXiv2025

CCJA: Context-Coherent Jailbreak Attack for Aligned Large Language Models

Despite explicit alignment efforts for large language models (LLMs), they can still be exploited to trigger unintended behaviors, a phenomenon known as "jailbreaking." Current jailbreak attack methods mainly focus on discrete prompt manipulations targeting closed-source LLMs, relying on manually crafted prompt templates and persuasion rules. However, as the capabilities of open-source LLMs improve, ensuring their safety becomes increasingly crucial. In such an environment, the accessibility of model parameters and gradient information by potential attackers exacerbates the severity of jailbreak threats. To address this research gap, we propose a novel \underline{C}ontext-\underline{C}oherent \underline{J}ailbreak \underline{A}ttack (CCJA). We define jailbreak attacks as an optimization problem within the embedding space of masked language models. Through combinatorial optimization, we effectively balance the jailbreak attack success rate with semantic coherence. Extensive evaluations show that our method not only maintains semantic consistency but also surpasses state-of-the-art baselines in attack effectiveness. Additionally, by integrating semantically coherent jailbreak prompts generated by our method into widely used black-box methodologies, we observe a notable enhancement in their success rates when targeting closed-source commercial LLMs. This highlights the security threat posed by open-source LLMs to commercial counterparts. We will open-source our code if the paper is accepted.

cs.CR↗

arXiv2025

On the Trustworthiness Landscape of State-of-the-art Generative Models: A Survey and Outlook

Diffusion models and large language models have emerged as leading-edge generative models, revolutionizing various aspects of human life. However, the practical implementations of these models have also exposed inherent risks, bringing to the forefront their evil sides and sparking concerns regarding their trustworthiness. Despite the wealth of literature on this subject, a comprehensive survey specifically delving into the intersection of large-scale generative models and their trustworthiness remains largely absent. To bridge this gap, this paper investigates both the long-standing and emerging threats associated with these models across four fundamental dimensions: 1) privacy, 2) security, 3) fairness, and 4) responsibility. Based on the investigation results, we develop an extensive map outlining the trustworthiness of large generative models. After that, we provide practical recommendations and potential research directions for future secure applications equipped with large generative models, ultimately promoting the trustworthiness of the models and benefiting the society as a whole.

cs.LG↗

arXiv2025

Bad-PFL: Exploring Backdoor Attacks against Personalized Federated Learning

Data heterogeneity and backdoor attacks rank among the most significant challenges facing federated learning (FL). For data heterogeneity, personalized federated learning (PFL) enables each client to maintain a private personalized model to cater to client-specific knowledge. Meanwhile, vanilla FL has proven vulnerable to backdoor attacks. However, recent advancements in PFL community have demonstrated a potential immunity against such attacks. This paper explores this intersection further, revealing that existing federated backdoor attacks fail in PFL because backdoors about manually designed triggers struggle to survive in personalized models. To tackle this, we design Bad-PFL, which employs features from natural data as our trigger. As long as the model is trained on natural data, it inevitably embeds the backdoor associated with our trigger, ensuring its longevity in personalized models. Moreover, our trigger undergoes mutual reinforcement training with the model, further solidifying the backdoor's durability and enhancing attack effectiveness. The large-scale experiments across three benchmark datasets demonstrate the superior performance of our attack against various PFL methods, even when equipped with state-of-the-art defense mechanisms.

cs.LG↗

arXiv2025

SVIA: A Street View Image Anonymization Framework for Self-Driving Applications

In recent years, there has been an increasing interest in image anonymization, particularly focusing on the de-identification of faces and individuals. However, for self-driving applications, merely de-identifying faces and individuals might not provide sufficient privacy protection since street views like vehicles and buildings can still disclose locations, trajectories, and other sensitive information. Therefore, it remains crucial to extend anonymization techniques to street view images to fully preserve the privacy of users, pedestrians, and vehicles. In this paper, we propose a Street View Image Anonymization (SVIA) framework for self-driving applications. The SVIA framework consists of three integral components: a semantic segmenter to segment an input image into functional regions, an inpainter to generate alternatives to privacy-sensitive regions, and a harmonizer to seamlessly stitch modified regions to guarantee visual coherence. Compared to existing methods, SVIA achieves a much better trade-off between image generation quality and privacy protection, as evidenced by experimental results for five common metrics on two widely used public datasets.

cs.CV↗

arXiv2025

Transferable Adversarial Examples with Bayes Approach

The vulnerability of deep neural networks (DNNs) to black-box adversarial attacks is one of the most heated topics in trustworthy AI. In such attacks, the attackers operate without any insider knowledge of the model, making the cross-model transferability of adversarial examples critical. Despite the potential for adversarial examples to be effective across various models, it has been observed that adversarial examples that are specifically crafted for a specific model often exhibit poor transferability. In this paper, we explore the transferability of adversarial examples via the lens of Bayesian approach. Specifically, we leverage Bayesian approach to probe the transferability and then study what constitutes a transferability-promoting prior. Following this, we design two concrete transferability-promoting priors, along with an adaptive dynamic weighting strategy for instances sampled from these priors. Employing these techniques, we present BayAtk. Extensive experiments illustrate the significant effectiveness of BayAtk in crafting more transferable adversarial examples against both undefended and defended black-box models compared to existing state-of-the-art attacks.

cs.LG↗

arXiv2024

ArtAug: Enhancing Text-to-Image Generation through Synthesis-Understanding Interaction

The emergence of diffusion models has significantly advanced image synthesis. The recent studies of model interaction and self-corrective reasoning approach in large language models offer new insights for enhancing text-to-image models. Inspired by these studies, we propose a novel method called ArtAug for enhancing text-to-image models in this paper. To the best of our knowledge, ArtAug is the first one that improves image synthesis models via model interactions with understanding models. In the interactions, we leverage human preferences implicitly learned by image understanding models to provide fine-grained suggestions for image synthesis models. The interactions can modify the image content to make it aesthetically pleasing, such as adjusting exposure, changing shooting angles, and adding atmospheric effects. The enhancements brought by the interaction are iteratively fused into the synthesis model itself through an additional enhancement module. This enables the synthesis model to directly produce aesthetically pleasing images without any extra computational cost. In the experiments, we train the ArtAug enhancement module on existing text-to-image models. Various evaluation metrics consistently demonstrate that ArtAug enhances the generative capabilities of text-to-image models without incurring additional computational costs. The source code and models will be released publicly.

cs.CV↗

arXiv2024

GenderAlign: An Alignment Dataset for Mitigating Gender Bias in Large Language Models

Large Language Models (LLMs) are prone to generating content that exhibits gender biases, raising significant ethical concerns. Alignment, the process of fine-tuning LLMs to better align with desired behaviors, is recognized as an effective approach to mitigate gender biases. Although proprietary LLMs have made significant strides in mitigating gender bias, their alignment datasets are not publicly available. The commonly used and publicly available alignment dataset, HH-RLHF, still exhibits gender bias to some extent. There is a lack of publicly available alignment datasets specifically designed to address gender bias. Hence, we developed a new dataset named GenderAlign, aiming at mitigating a comprehensive set of gender biases in LLMs. This dataset comprises 8k single-turn dialogues, each paired with a "chosen" and a "rejected" response. Compared to the "rejected" responses, the "chosen" responses demonstrate lower levels of gender bias and higher quality. Furthermore, we categorized the gender biases in the "rejected" responses of GenderAlign into 4 principal categories. The experimental results show the effectiveness of GenderAlign in reducing gender bias in LLMs.

cs.CL↗