With the ability to generate high-quality images, text-to-image (T2I) models can be exploited for creating inappropriate content. To prevent misuse, existing safety measures are either based on text blacklists, which can be easily circumvented, or harmful content classification, requiring large datasets for training and offering low flexibility. Hence, we propose Latent Guard, a framework designed to improve safety measures in text-to-image generation. Inspired by blacklist-based approaches, Latent Guard learns a latent space on top of the T2I model's text encoder, where it is possible to check the presence of harmful concepts in the input text embeddings. Our proposed framework is composed of a data generation pipeline specific to the task using large language models, ad-hoc architectural components, and a contrastive learning strategy to benefit from the generated data. The effectiveness of our method is verified on three datasets and against four baselines. Code and data will be shared at https://github.com/rt219/LatentGuard.
In recent years, generative AI (GenAI), like large language models and text-to-image models, has received significant attention across various domains. However, ensuring the responsible generation of content by these models is crucial for their real-world applicability. This raises an interesting question: \textit{What should responsible GenAI generate, and what should it not?} To answer the question, this paper investigates the practical responsible requirements of both textual and visual generative models, outlining five key considerations: generating truthful content, avoiding toxic content, refusing harmful instruction, leaking no training data-related content, and ensuring generated content identifiable. Specifically, we review recent advancements and challenges in addressing these requirements. Besides, we discuss and emphasize the importance of responsible GenAI across healthcare, education, finance, and artificial general intelligence domains. Through a unified perspective on both textual and visual generative models, this paper aims to provide insights into practical safety-related issues and further benefit the community in building responsible GenAI.
Various jailbreak attacks have been proposed to red-team Large Language Models (LLMs) and revealed the vulnerable safeguards of LLMs. Besides, some methods are not limited to the textual modality and extend the jailbreak attack to Multimodal Large Language Models (MLLMs) by perturbing the visual input. However, the absence of a universal evaluation benchmark complicates the performance reproduction and fair comparison. Besides, there is a lack of comprehensive evaluation of closed-source state-of-the-art (SOTA) models, especially MLLMs, such as GPT-4V. To address these issues, this work first builds a comprehensive jailbreak evaluation dataset with 1445 harmful questions covering 11 different safety policies. Based on this dataset, extensive red-teaming experiments are conducted on 11 different LLMs and MLLMs, including both SOTA proprietary models and open-source models. We then conduct a deep analysis of the evaluated results and find that (1) GPT4 and GPT-4V demonstrate better robustness against jailbreak attacks compared to open-source LLMs and MLLMs. (2) Llama2 and Qwen-VL-Chat are more robust compared to other open-source models. (3) The transferability of visual jailbreak methods is relatively limited compared to textual jailbreak methods. The dataset and code can be found here https://anonymous.4open.science/r/red_teaming_gpt4-C1CE/README.md .
Recent progress in visual generative models enables the generation of high-quality images. To prevent the misuse of generated images, it is important to identify the origin model that generates them. In this work, we study the origin attribution of generated images in a practical setting where only a few images generated by a source model are available and the source model cannot be accessed. The goal is to check if a given image is generated by the source model. We first formulate this problem as a few-shot one-class classification task. To solve the task, we propose OCC-CLIP, a CLIP-based framework for few-shot one-class classification, enabling the identification of an image's source model, even among multiple candidates. Extensive experiments corresponding to various generative models verify the effectiveness of our OCC-CLIP framework. Furthermore, an experiment based on the recently released DALL-E 3 API verifies the real-world applicability of our solution.
Foundation models pre-trained on web-scale vision-language data, such as CLIP, are widely used as cornerstones of powerful machine learning systems. While pre-training offers clear advantages for downstream learning, it also endows downstream models with shared adversarial vulnerabilities that can be easily identified through the open-sourced foundation model. In this work, we expose such vulnerabilities in CLIP's downstream models and show that foundation models can serve as a basis for attacking their downstream systems. In particular, we propose a simple yet effective adversarial attack strategy termed Patch Representation Misalignment (PRM). Solely based on open-sourced CLIP vision encoders, this method produces adversaries that simultaneously fool more than 20 downstream models spanning 4 common vision-language tasks (semantic segmentation, object detection, image captioning and visual question-answering). Our findings highlight the concerning safety risks introduced by the extensive usage of public foundational models in the development of downstream systems, calling for extra caution in these scenarios.
Different from traditional task-specific vision models, recent large VLMs can readily adapt to different vision tasks by simply using different textual instructions, i.e., prompts. However, a well-known concern about traditional task-specific vision models is that they can be misled by imperceptible adversarial perturbations. Furthermore, the concern is exacerbated by the phenomenon that the same adversarial perturbations can fool different task-specific models. Given that VLMs rely on prompts to adapt to different tasks, an intriguing question emerges: Can a single adversarial image mislead all predictions of VLMs when a thousand different prompts are given? This question essentially introduces a novel perspective on adversarial transferability: cross-prompt adversarial transferability. In this work, we propose the Cross-Prompt Attack (CroPA). This proposed method updates the visual adversarial perturbation with learnable prompts, which are designed to counteract the misleading effects of the adversarial image. By doing this, CroPA significantly improves the transferability of adversarial examples across prompts. Extensive experiments are conducted to verify the strong cross-prompt adversarial transferability of CroPA with prevalent VLMs including Flamingo, BLIP-2, and InstructBLIP in various different tasks. Our source code is available at \url{https://github.com/Haochen-Luo/CroPA}.
Adversarial attack methods based on point manipulation for 3D point cloud classification have revealed the fragility of 3D models, yet the adversarial examples they produce are easily perceived or defended against. The trade-off between the imperceptibility and adversarial strength leads most point attack methods to inevitably introduce easily detectable outlier points upon a successful attack. Another promising strategy, shape-based attack, can effectively eliminate outliers, but existing methods often suffer significant reductions in imperceptibility due to irrational deformations. We find that concealing deformation perturbations in areas insensitive to human eyes can achieve a better trade-off between imperceptibility and adversarial strength, specifically in parts of the object surface that are complex and exhibit drastic curvature changes. Therefore, we propose a novel shape-based adversarial attack method, HiT-ADV, which initially conducts a two-stage search for attack regions based on saliency and imperceptibility scores, and then adds deformation perturbations in each attack region using Gaussian kernel functions. Additionally, HiT-ADV is extendable to physical attack. We propose that by employing benign resampling and benign rigid transformations, we can further enhance physical adversarial strength with little sacrifice to imperceptibility. Extensive experiments have validated the superiority of our method in terms of adversarial and imperceptible properties in both digital and physical spaces. Our code is avaliable at: https://github.com/TRLou/HiT-ADV.
Recently, Multimodal LLMs (MLLMs) have shown a great ability to understand images. However, like traditional vision models, they are still vulnerable to adversarial images. Meanwhile, Chain-of-Thought (CoT) reasoning has been widely explored on MLLMs, which not only improves model's performance, but also enhances model's explainability by giving intermediate reasoning steps. Nevertheless, there is still a lack of study regarding MLLMs' adversarial robustness with CoT and an understanding of what the rationale looks like when MLLMs infer wrong answers with adversarial images. Our research evaluates the adversarial robustness of MLLMs when employing CoT reasoning, finding that CoT marginally improves adversarial robustness against existing attack methods. Moreover, we introduce a novel stop-reasoning attack technique that effectively bypasses the CoT-induced robustness enhancements. Finally, we demonstrate the alterations in CoT reasoning when MLLMs confront adversarial images, shedding light on their reasoning process under adversarial attacks.
Large vision-language models (VLMs) such as GPT-4 have achieved exceptional performance across various multi-modal tasks. However, the deployment of VLMs necessitates substantial energy consumption and computational resources. Once attackers maliciously induce high energy consumption and latency time (energy-latency cost) during inference of VLMs, it will exhaust computational resources. In this paper, we explore this attack surface about availability of VLMs and aim to induce high energy-latency cost during inference of VLMs. We find that high energy-latency cost during inference of VLMs can be manipulated by maximizing the length of generated sequences. To this end, we propose verbose images, with the goal of crafting an imperceptible perturbation to induce VLMs to generate long sentences during inference. Concretely, we design three loss objectives. First, a loss is proposed to delay the occurrence of end-of-sequence (EOS) token, where EOS token is a signal for VLMs to stop generating further tokens. Moreover, an uncertainty loss and a token diversity loss are proposed to increase the uncertainty over each generated token and the diversity among all tokens of the whole generated sequence, respectively, which can break output dependency at token-level and sequence-level. Furthermore, a temporal weight adjustment algorithm is proposed, which can effectively balance these losses. Extensive experiments demonstrate that our verbose images can increase the length of generated sequences by 7.87 times and 8.56 times compared to original images on MS-COCO and ImageNet datasets, which presents potential challenges for various applications. Our code is available at https://github.com/KuofengGao/Verbose_Images.
The field of few-shot learning (FSL) has shown promising results in scenarios where training data is limited, but its vulnerability to backdoor attacks remains largely unexplored. We first explore this topic by first evaluating the performance of the existing backdoor attack methods on few-shot learning scenarios. Unlike in standard supervised learning, existing backdoor attack methods failed to perform an effective attack in FSL due to two main issues. Firstly, the model tends to overfit to either benign features or trigger features, causing a tough trade-off between attack success rate and benign accuracy. Secondly, due to the small number of training samples, the dirty label or visible trigger in the support set can be easily detected by victims, which reduces the stealthiness of attacks. It seemed that FSL could survive from backdoor attacks. However, in this paper, we propose the Few-shot Learning Backdoor Attack (FLBA) to show that FSL can still be vulnerable to backdoor attacks. Specifically, we first generate a trigger to maximize the gap between poisoned and benign features. It enables the model to learn both benign and trigger features, which solves the problem of overfitting. To make it more stealthy, we hide the trigger by optimizing two types of imperceptible perturbation, namely attractive and repulsive perturbation, instead of attaching the trigger directly. Once we obtain the perturbations, we can poison all samples in the benign support set into a hidden poisoned support set and fine-tune the model on it. Our method demonstrates a high Attack Success Rate (ASR) in FSL tasks with different few-shot learning paradigms while preserving clean accuracy and maintaining stealthiness. This study reveals that few-shot learning still suffers from backdoor attacks, and its security should be given attention.