Search

Masked Image Modeling (MIM) has achieved significant success in the realm of self-supervised learning (SSL) for visual recognition. The image encoder pre-trained through MIM, involving the masking and subsequent reconstruction of input images, attains state-of-the-art performance in various downstream vision tasks. However, most existing works focus on improving the performance of MIM.In this work, we take a different angle by studying the pre-training data privacy of MIM. Specifically, we propose the first membership inference attack against image encoders pre-trained by MIM, which aims to determine whether an image is part of the MIM pre-training dataset. The key design is to simulate the pre-training paradigm of MIM, i.e., image masking and subsequent reconstruction, and then obtain reconstruction errors. These reconstruction errors can serve as membership signals for achieving attack goals, as the encoder is more capable of reconstructing the input image in its training set with lower errors. Extensive evaluations are conducted on three model architectures and three benchmark datasets. Empirical results show that our attack outperforms baseline methods. Additionally, we undertake intricate ablation studies to analyze multiple factors that could influence the performance of the attack.

Large Language Models (LLMs) excel in various applications, including text generation and complex tasks. However, the misuse of LLMs raises concerns about the authenticity and ethical implications of the content they produce, such as deepfake news, academic fraud, and copyright infringement. Watermarking techniques, which embed identifiable markers in machine-generated text, offer a promising solution to these issues by allowing for content verification and origin tracing. Unfortunately, the robustness of current LLM watermarking schemes under potential watermark removal attacks has not been comprehensively explored. In this paper, to fill this gap, we first systematically comb the mainstream watermarking schemes and removal attacks on machine-generated texts, and then we categorize them into pre-text (before text generation) and post-text (after text generation) classes so that we can conduct diversified analyses. In our experiments, we evaluate eight watermarks (five pre-text, three post-text) and twelve attacks (two pre-text, ten post-text) across 87 scenarios. Evaluation results indicate that (1) KGW and Exponential watermarks offer high text quality and watermark retention but remain vulnerable to most attacks; (2) Post-text attacks are found to be more efficient and practical than pre-text attacks; (3) Pre-text watermarks are generally more imperceptible, as they do not alter text fluency, unlike post-text watermarks; (4) Additionally, combined attack methods can significantly increase effectiveness, highlighting the need for more robust watermarking solutions. Our study underscores the vulnerabilities of current techniques and the necessity for developing more resilient schemes.

Large Language Models (LLMs) have performed exceptionally in various text-generative tasks, including question answering, translation, code completion, etc. However, the over-assistance of LLMs has raised the challenge of "jailbreaking", which induces the model to generate malicious responses against the usage policy and society by designing adversarial prompts. With the emergence of jailbreak attack methods exploiting different vulnerabilities in LLMs, the corresponding safety alignment measures are also evolving. In this paper, we propose a comprehensive and detailed taxonomy of jailbreak attack and defense methods. For instance, the attack methods are divided into black-box and white-box attacks based on the transparency of the target model. Meanwhile, we classify defense methods into prompt-level and model-level defenses. Additionally, we further subdivide these attack and defense methods into distinct sub-classes and present a coherent diagram illustrating their relationships. We also conduct an investigation into the current evaluation methods and compare them from different perspectives. Our findings aim to inspire future research and practical implementations in safeguarding LLMs against adversarial attacks. Above all, although jailbreak remains a significant concern within the community, we believe that our work enhances the understanding of this domain and provides a foundation for developing more secure LLMs.

Jailbreak attacks aim to induce Large Language Models (LLMs) to generate harmful responses for forbidden instructions, presenting severe misuse threats to LLMs. Up to now, research into jailbreak attacks and defenses is emerging, however, there is (surprisingly) no consensus on how to evaluate whether a jailbreak attempt is successful. In other words, the methods to assess the harmfulness of an LLM's response are varied, such as manual annotation or prompting GPT-4 in specific ways. Each approach has its own set of strengths and weaknesses, impacting their alignment with human values, as well as the time and financial cost. This diversity in evaluation presents challenges for researchers in choosing suitable evaluation methods and conducting fair comparisons across different jailbreak attacks and defenses. In this paper, we conduct a comprehensive analysis of jailbreak evaluation methodologies, drawing from nearly ninety jailbreak research released between May 2023 and April 2024. Our study introduces a systematic taxonomy of jailbreak evaluators, offering in-depth insights into their strengths and weaknesses, along with the current status of their adaptation. Moreover, to facilitate subsequent research, we propose JailbreakEval, a user-friendly toolkit focusing on the evaluation of jailbreak attempts. It includes various well-known evaluators out-of-the-box, so that users can obtain evaluation results with only a single command. JailbreakEval also allows users to customize their own evaluation workflow in a unified framework with the ease of development and comparison. In summary, we regard JailbreakEval to be a catalyst that simplifies the evaluation process in jailbreak research and fosters an inclusive standard for jailbreak evaluation within the community., Comment: Our code is available at https://github.com/ThuCCSLab/JailbreakEval

Despite the remarkable advance of large language models (LLMs), the prevalence of non-factual responses remains a common issue. This work studies non-factuality prediction (NFP), which predicts whether an LLM will generate non-factual responses to a question before the generation process. Previous efforts on NFP usually rely on extensive computation. In this work, we conduct extensive analysis to explore the capabilities of using a lightweight probe to elicit ``whether an LLM knows'' from the hidden representations of questions. Additionally, we discover that the non-factuality probe employs similar patterns for NFP across multiple LLMs. Motivated by the intriguing finding, we conduct effective transfer learning for cross-LLM NFP and propose a question-aligned strategy to ensure the efficacy of mini-batch based training.

A graph neural network (GNN) is a type of neural network that is specifically designed to process graph-structured data. Typically, GNNs can be implemented in two settings, including the transductive setting and the inductive setting. In the transductive setting, the trained model can only predict the labels of nodes that were observed at the training time. In the inductive setting, the trained model can be generalized to new nodes/graphs. Due to its flexibility, the inductive setting is the most popular GNN setting at the moment. Previous work has shown that transductive GNNs are vulnerable to a series of privacy attacks. However, a comprehensive privacy analysis of inductive GNN models is still missing. This paper fills the gap by conducting a systematic privacy analysis of inductive GNNs through the lens of link stealing attacks, one of the most popular attacks that are specifically designed for GNNs. We propose two types of link stealing attacks, i.e., posterior-only attacks and combined attacks. We define threat models of the posterior-only attacks with respect to node topology and the combined attacks by considering combinations of posteriors, node attributes, and graph features. Extensive evaluation on six real-world datasets demonstrates that inductive GNNs leak rich information that enables link stealing attacks with advantageous properties. Even attacks with no knowledge about graph structures can be effective. We also show that our attacks are robust to different node similarities and different graph features. As a counterpart, we investigate two possible defenses and discover they are ineffective against our attacks, which calls for more effective defenses., Comment: To appear in the 24th Privacy Enhancing Technologies Symposium (PETS 2024), July 15-20, 2024

Model merging is a promising lightweight model empowerment technique that does not rely on expensive computing devices (e.g., GPUs) or require the collection of specific training data. Instead, it involves editing different upstream model parameters to absorb their downstream task capabilities. However, uncertified model merging can infringe upon the Intellectual Property (IP) rights of the original upstream models. In this paper, we conduct the first study on the robustness of IP protection methods in model merging scenarios. We investigate two state-of-the-art IP protection techniques: Quantization Watermarking and Instructional Fingerprint, along with various advanced model merging technologies, such as Task Arithmetic, TIES-MERGING, and so on. Experimental results indicate that current Large Language Model (LLM) watermarking techniques cannot survive in the merged models, whereas model fingerprinting techniques can. Our research aims to highlight that model merging should be an indispensable consideration in the robustness assessment of model IP protection techniques, thereby promoting the healthy development of the open-source LLM community., Comment: Technical Report

While advanced machine learning (ML) models are deployed in numerous real-world applications, previous works demonstrate these models have security and privacy vulnerabilities. Various empirical research has been done in this field. However, most of the experiments are performed on target ML models trained by the security researchers themselves. Due to the high computational resource requirement for training advanced models with complex architectures, researchers generally choose to train a few target models using relatively simple architectures on typical experiment datasets. We argue that to understand ML models' vulnerabilities comprehensively, experiments should be performed on a large set of models trained with various purposes (not just the purpose of evaluating ML attacks and defenses). To this end, we propose using publicly available models with weights from the Internet (public models) for evaluating attacks and defenses on ML models. We establish a database, namely SecurityNet, containing 910 annotated image classification models. We then analyze the effectiveness of several representative attacks/defenses, including model stealing attacks, membership inference attacks, and backdoor detection on these public models. Our evaluation empirically shows the performance of these attacks/defenses can vary significantly on public models compared to self-trained models. We share SecurityNet with the research community. and advocate researchers to perform experiments on public models to better demonstrate their proposed methods' effectiveness in the future., Comment: To appear in the 33rd USENIX Security Symposium, August 2024, Philadelphia, PA, USA

Training a machine learning model with data following a meaningful order, i.e., from easy to hard, has been proven to be effective in accelerating the training process and achieving better model performance. The key enabling technique is curriculum learning (CL), which has seen great success and has been deployed in areas like image and text classification. Yet, how CL affects the privacy of machine learning is unclear. Given that CL changes the way a model memorizes the training data, its influence on data privacy needs to be thoroughly evaluated. To fill this knowledge gap, we perform the first study and leverage membership inference attack (MIA) and attribute inference attack (AIA) as two vectors to quantify the privacy leakage caused by CL. Our evaluation of nine real-world datasets with attack methods (NN-based, metric-based, label-only MIA, and NN-based AIA) revealed new insights about CL. First, MIA becomes slightly more effective when CL is applied, but the impact is much more prominent to a subset of training samples ranked as difficult. Second, a model trained under CL is less vulnerable under AIA, compared to MIA. Third, the existing defense techniques like DP-SGD, MemGuard, and MixupMMD are still effective under CL, though DP-SGD has a significant impact on target model accuracy. Finally, based on our insights into CL, we propose a new MIA, termed Diff-Cali, which exploits the difficulty scores for result calibration and is demonstrated to be effective against all CL methods and the normal training method. With this study, we hope to draw the community's attention to the unintended privacy risks of emerging machine-learning techniques and develop new attack benchmarks and defense solutions.

Deploying machine learning (ML) models in the wild is challenging as it suffers from distribution shifts, where the model trained on an original domain cannot generalize well to unforeseen diverse transfer domains. To address this challenge, several test-time adaptation (TTA) methods have been proposed to improve the generalization ability of the target pre-trained models under test data to cope with the shifted distribution. The success of TTA can be credited to the continuous fine-tuning of the target model according to the distributional hint from the test samples during test time. Despite being powerful, it also opens a new attack surface, i.e., test-time poisoning attacks, which are substantially different from previous poisoning attacks that occur during the training time of ML models (i.e., adversaries cannot intervene in the training process). In this paper, we perform the first test-time poisoning attack against four mainstream TTA methods, including TTT, DUA, TENT, and RPL. Concretely, we generate poisoned samples based on the surrogate models and feed them to the target TTA models. Experimental results show that the TTA methods are generally vulnerable to test-time poisoning attacks. For instance, the adversary can feed as few as 10 poisoned samples to degrade the performance of the target model from 76.20% to 41.83%. Our results demonstrate that TTA algorithms lacking a rigorous security assessment are unsuitable for deployment in real-life scenarios. As such, we advocate for the integration of defenses against test-time poisoning attacks into the design of TTA methods., Comment: To Appear in the 45th IEEE Symposium on Security and Privacy, May 20-23, 2024

The spread of toxic content online is an important problem that has adverse effects on user experience online and in our society at large. Motivated by the importance and impact of the problem, research focuses on developing solutions to detect toxic content, usually leveraging machine learning (ML) models trained on human-annotated datasets. While these efforts are important, these models usually do not generalize well and they can not cope with new trends (e.g., the emergence of new toxic terms). Currently, we are witnessing a shift in the approach to tackling societal issues online, particularly leveraging large language models (LLMs) like GPT-3 or T5 that are trained on vast corpora and have strong generalizability. In this work, we investigate how we can use LLMs and prompt learning to tackle the problem of toxic content, particularly focusing on three tasks; 1) Toxicity Classification, 2) Toxic Span Detection, and 3) Detoxification. We perform an extensive evaluation over five model architectures and eight datasets demonstrating that LLMs with prompt learning can achieve similar or even better performance compared to models trained on these specific tasks. We find that prompt learning achieves around 10\% improvement in the toxicity classification task compared to the baselines, while for the toxic span detection task we find better performance to the best baseline (0.643 vs. 0.640 in terms of $F_1$-score). Finally, for the detoxification task, we find that prompt learning can successfully reduce the average toxicity score (from 0.775 to 0.213) while preserving semantic meaning., Comment: To Appear in the 45th IEEE Symposium on Security and Privacy, May 20-23, 2024

Graph generative models become increasingly effective for data distribution approximation and data augmentation. While they have aroused public concerns about their malicious misuses or misinformation broadcasts, just as what Deepfake visual and auditory media has been delivering to society. Hence it is essential to regulate the prevalence of generated graphs. To tackle this problem, we pioneer the formulation of the generated graph detection problem to distinguish generated graphs from real ones. We propose the first framework to systematically investigate a set of sophisticated models and their performance in four classification scenarios. Each scenario switches between seen and unseen datasets/generators during testing to get closer to real-world settings and progressively challenge the classifiers. Extensive experiments evidence that all the models are qualified for generated graph detection, with specific models having advantages in specific scenarios. Resulting from the validated generality and oblivion of the classifiers to unseen datasets/generators, we draw a safe conclusion that our solution can sustain for a decent while to curb generated graph misuses., Comment: Accepted by ICML 2023

Large text-to-image models have shown remarkable performance in synthesizing high-quality images. In particular, the subject-driven model makes it possible to personalize the image synthesis for a specific subject, e.g., a human face or an artistic style, by fine-tuning the generic text-to-image model with a few images from that subject. Nevertheless, misuse of subject-driven image synthesis may violate the authority of subject owners. For example, malicious users may use subject-driven synthesis to mimic specific artistic styles or to create fake facial images without authorization. To protect subject owners against such misuse, recent attempts have commonly relied on adversarial examples to indiscriminately disrupt subject-driven image synthesis. However, this essentially prevents any benign use of subject-driven synthesis based on protected images. In this paper, we take a different angle and aim at protection without sacrificing the utility of protected images for general synthesis purposes. Specifically, we propose GenWatermark, a novel watermark system based on jointly learning a watermark generator and a detector. In particular, to help the watermark survive the subject-driven synthesis, we incorporate the synthesis process in learning GenWatermark by fine-tuning the detector with synthesized images for a specific subject. This operation is shown to largely improve the watermark detection accuracy and also ensure the uniqueness of the watermark for each individual subject. Extensive experiments validate the effectiveness of GenWatermark, especially in practical scenarios with unknown models and text prompts (74% Acc.), as well as partial data watermarking (80% Acc. for 1/4 watermarking). We also demonstrate the robustness of GenWatermark to two potential countermeasures that substantially degrade the synthesis quality.

State-of-the-art Text-to-Image models like Stable Diffusion and DALLE$\cdot$2 are revolutionizing how people generate visual content. At the same time, society has serious concerns about how adversaries can exploit such models to generate unsafe images. In this work, we focus on demystifying the generation of unsafe images and hateful memes from Text-to-Image models. We first construct a typology of unsafe images consisting of five categories (sexually explicit, violent, disturbing, hateful, and political). Then, we assess the proportion of unsafe images generated by four advanced Text-to-Image models using four prompt datasets. We find that these models can generate a substantial percentage of unsafe images; across four models and four prompt datasets, 14.56% of all generated images are unsafe. When comparing the four models, we find different risk levels, with Stable Diffusion being the most prone to generating unsafe content (18.92% of all generated images are unsafe). Given Stable Diffusion's tendency to generate more unsafe content, we evaluate its potential to generate hateful meme variants if exploited by an adversary to attack a specific individual or community. We employ three image editing methods, DreamBooth, Textual Inversion, and SDEdit, which are supported by Stable Diffusion. Our evaluation result shows that 24% of the generated images using DreamBooth are hateful meme variants that present the features of the original hateful meme and the target individual/community; these generated images are comparable to hateful meme variants collected from the real world. Overall, our results demonstrate that the danger of large-scale generation of unsafe images is imminent. We discuss several mitigating measures, such as curating training data, regulating prompts, and implementing safety filters, and encourage better safeguard tools to be developed to prevent unsafe generation., Comment: To Appear in the ACM Conference on Computer and Communications Security, November 26, 2023

Nowadays, powerful large language models (LLMs) such as ChatGPT have demonstrated revolutionary power in a variety of tasks. Consequently, the detection of machine-generated texts (MGTs) is becoming increasingly crucial as LLMs become more advanced and prevalent. These models have the ability to generate human-like language, making it challenging to discern whether a text is authored by a human or a machine. This raises concerns regarding authenticity, accountability, and potential bias. However, existing methods for detecting MGTs are evaluated using different model architectures, datasets, and experimental settings, resulting in a lack of a comprehensive evaluation framework that encompasses various methodologies. Furthermore, it remains unclear how existing detection methods would perform against powerful LLMs. In this paper, we fill this gap by proposing the first benchmark framework for MGT detection against powerful LLMs, named MGTBench. Extensive evaluations on public datasets with curated texts generated by various powerful LLMs such as ChatGPT-turbo and Claude demonstrate the effectiveness of different detection methods. Our ablation study shows that a larger number of words in general leads to better performance and most detection methods can achieve similar performance with much fewer training samples. Moreover, we delve into a more challenging task: text attribution. Our findings indicate that the model-based detection methods still perform well in the text attribution task. To investigate the robustness of different detection methods, we consider three adversarial attacks, namely paraphrasing, random spacing, and adversarial perturbations. We discover that these attacks can significantly diminish detection effectiveness, underscoring the critical need for the development of more robust detection methods.

Building advanced machine learning (ML) models requires expert knowledge and many trials to discover the best architecture and hyperparameter settings. Previous work demonstrates that model information can be leveraged to assist other attacks, such as membership inference, generating adversarial examples. Therefore, such information, e.g., hyperparameters, should be kept confidential. It is well known that an adversary can leverage a target ML model's output to steal the model's information. In this paper, we discover a new side channel for model information stealing attacks, i.e., models' scientific plots which are extensively used to demonstrate model performance and are easily accessible. Our attack is simple and straightforward. We leverage the shadow model training techniques to generate training data for the attack model which is essentially an image classifier. Extensive evaluation on three benchmark datasets shows that our proposed attack can effectively infer the architecture/hyperparameters of image classifiers based on convolutional neural network (CNN) given the scientific plot generated from it. We also reveal that the attack's success is mainly caused by the shape of the scientific plots, and further demonstrate that the attacks are robust in various scenarios. Given the simplicity and effectiveness of the attack method, our study indicates scientific plots indeed constitute a valid side channel for model information stealing attacks. To mitigate the attacks, we propose several defense mechanisms that can reduce the original attacks' accuracy while maintaining the plot utility. However, such defenses can still be bypassed by adaptive attacks., Comment: To appear in the 32nd USENIX Security Symposium, August 2023, Anaheim, CA, USA

Backdoor attacks represent one of the major threats to machine learning models. Various efforts have been made to mitigate backdoors. However, existing defenses have become increasingly complex and often require high computational resources or may also jeopardize models' utility. In this work, we show that fine-tuning, one of the most common and easy-to-adopt machine learning training operations, can effectively remove backdoors from machine learning models while maintaining high model utility. Extensive experiments over three machine learning paradigms show that fine-tuning and our newly proposed super-fine-tuning achieve strong defense performance. Furthermore, we coin a new term, namely backdoor sequela, to measure the changes in model vulnerabilities to other attacks before and after the backdoor has been removed. Empirical evaluation shows that, compared to other defense methods, super-fine-tuning leaves limited backdoor sequela. We hope our results can help machine learning model owners better protect their models from backdoor threats. Also, it calls for the design of more advanced attacks in order to comprehensively assess machine learning models' backdoor vulnerabilities., Comment: 17 pages, 17 figures

The dissemination of hateful memes online has adverse effects on social media platforms and the real world. Detecting hateful memes is challenging, one of the reasons being the evolutionary nature of memes; new hateful memes can emerge by fusing hateful connotations with other cultural ideas or symbols. In this paper, we propose a framework that leverages multimodal contrastive learning models, in particular OpenAI's CLIP, to identify targets of hateful content and systematically investigate the evolution of hateful memes. We find that semantic regularities exist in CLIP-generated embeddings that describe semantic relationships within the same modality (images) or across modalities (images and text). Leveraging this property, we study how hateful memes are created by combining visual elements from multiple images or fusing textual information with a hateful image. We demonstrate the capabilities of our framework for analyzing the evolution of hateful memes by focusing on antisemitic memes, particularly the Happy Merchant meme. Using our framework on a dataset extracted from 4chan, we find 3.3K variants of the Happy Merchant meme, with some linked to specific countries, persons, or organizations. We envision that our framework can be used to aid human moderators by flagging new variants of hateful memes so that moderators can manually verify them and mitigate the problem of hateful content online., Comment: To Appear in the 44th IEEE Symposium on Security and Privacy, May 22-25, 2023

Masked image modeling (MIM) revolutionizes self-supervised learning (SSL) for image pre-training. In contrast to previous dominating self-supervised methods, i.e., contrastive learning, MIM attains state-of-the-art performance by masking and reconstructing random patches of the input image. However, the associated security and privacy risks of this novel generative method are unexplored. In this paper, we perform the first security risk quantification of MIM through the lens of backdoor attacks. Different from previous work, we are the first to systematically threat modeling on SSL in every phase of the model supply chain, i.e., pre-training, release, and downstream phases. Our evaluation shows that models built with MIM are vulnerable to existing backdoor attacks in release and downstream phases and are compromised by our proposed method in pre-training phase. For instance, on CIFAR10, the attack success rate can reach 99.62%, 96.48%, and 98.89% in the downstream phase, release phase, and pre-training phase, respectively. We also take the first step to investigate the success factors of backdoor attacks in the pre-training phase and find the trigger number and trigger pattern play key roles in the success of backdoor attacks while trigger location has only tiny effects. In the end, our empirical study of the defense mechanisms across three detection-level on model supply chain phases indicates that different defenses are suitable for backdoor attacks in different phases. However, backdoor attacks in the release phase cannot be detected by all three detection-level methods, calling for more effective defenses in future research.

Recently, the newly emerged multimodal models, which leverage both visual and linguistic modalities to train powerful encoders, have gained increasing attention. However, learning from a large-scale unlabeled dataset also exposes the model to the risk of potential poisoning attacks, whereby the adversary aims to perturb the model's training data to trigger malicious behaviors in it. In contrast to previous work, only poisoning visual modality, in this work, we take the first step to studying poisoning attacks against multimodal models in both visual and linguistic modalities. Specially, we focus on answering two questions: (1) Is the linguistic modality also vulnerable to poisoning attacks? and (2) Which modality is most vulnerable? To answer the two questions, we propose three types of poisoning attacks against multimodal models. Extensive evaluations on different datasets and model architectures show that all three attacks can achieve significant attack performance while maintaining model utility in both visual and linguistic modalities. Furthermore, we observe that the poisoning effect differs between different modalities. To mitigate the attacks, we propose both pre-training and post-training defenses. We empirically show that both defenses can significantly reduce the attack performance while preserving the model's utility., Comment: To Appear in the 40th International Conference on Machine Learning, July 2023

Relying on the fact that not all inputs require the same amount of computation to yield a confident prediction, multi-exit networks are gaining attention as a prominent approach for pushing the limits of efficient deployment. Multi-exit networks endow a backbone model with early exits, allowing to obtain predictions at intermediate layers of the model and thus save computation time and/or energy. However, current various designs of multi-exit networks are only considered to achieve the best trade-off between resource usage efficiency and prediction accuracy, the privacy risks stemming from them have never been explored. This prompts the need for a comprehensive investigation of privacy risks in multi-exit networks. In this paper, we perform the first privacy analysis of multi-exit networks through the lens of membership leakages. In particular, we first leverage the existing attack methodologies to quantify the multi-exit networks' vulnerability to membership leakages. Our experimental results show that multi-exit networks are less vulnerable to membership leakages and the exit (number and depth) attached to the backbone model is highly correlated with the attack performance. Furthermore, we propose a hybrid attack that exploits the exit information to improve the performance of existing attacks. We evaluate membership leakage threat caused by our hybrid attack under three different adversarial setups, ultimately arriving at a model-free and data-free adversary. These results clearly demonstrate that our hybrid attacks are very broadly applicable, thereby the corresponding risks are much more severe than shown by existing membership inference attacks. We further present a defense mechanism called TimeGuard specifically for multi-exit networks and show that TimeGuard mitigates the newly proposed attacks perfectly., Comment: Accepted by CCS 2022

Machine learning models are prone to memorizing sensitive data, making them vulnerable to membership inference attacks in which an adversary aims to infer whether an input sample was used to train the model. Over the past few years, researchers have produced many membership inference attacks and defenses. However, these attacks and defenses employ a variety of strategies and are conducted in different models and datasets. The lack of comprehensive benchmark, however, means we do not understand the strengths and weaknesses of existing attacks and defenses. We fill this gap by presenting a large-scale measurement of different membership inference attacks and defenses. We systematize membership inference through the study of nine attacks and six defenses and measure the performance of different attacks and defenses in the holistic evaluation. We then quantify the impact of the threat model on the results of these attacks. We find that some assumptions of the threat model, such as same-architecture and same-distribution between shadow and target models, are unnecessary. We are also the first to execute attacks on the real-world data collected from the Internet, instead of laboratory datasets. We further investigate what determines the performance of membership inference attacks and reveal that the commonly believed overfitting level is not sufficient for the success of the attacks. Instead, the Jensen-Shannon distance of entropy/cross-entropy between member and non-member samples correlates with attack performance much better. This gives us a new way to accurately predict membership inference risks without running the attack. Finally, we find that data augmentation degrades the performance of existing attacks to a larger extent, and we propose an adaptive attack using augmentation to train shadow and attack models that improve attack performance.

Semi-supervised learning (SSL) leverages both labeled and unlabeled data to train machine learning (ML) models. State-of-the-art SSL methods can achieve comparable performance to supervised learning by leveraging much fewer labeled data. However, most existing works focus on improving the performance of SSL. In this work, we take a different angle by studying the training data privacy of SSL. Specifically, we propose the first data augmentation-based membership inference attacks against ML models trained by SSL. Given a data sample and the black-box access to a model, the goal of membership inference attack is to determine whether the data sample belongs to the training dataset of the model. Our evaluation shows that the proposed attack can consistently outperform existing membership inference attacks and achieves the best performance against the model trained by SSL. Moreover, we uncover that the reason for membership leakage in SSL is different from the commonly believed one in supervised learning, i.e., overfitting (the gap between training and testing accuracy). We observe that the SSL model is well generalized to the testing data (with almost 0 overfitting) but ''memorizes'' the training data by giving a more confident prediction regardless of its correctness. We also explore early stopping as a countermeasure to prevent membership inference attacks against SSL. The results show that early stopping can mitigate the membership inference attack, but with the cost of model's utility degradation., Comment: Accepted to ECCV 2022

Sinophobia, anti-Chinese sentiment, has existed on the Web for a long time. The outbreak of COVID-19 and the extended quarantine has further amplified it. However, we lack a quantitative understanding of the cause of Sinophobia as well as how it evolves over time. In this paper, we conduct a large-scale longitudinal measurement of Sinophobia, between 2016 and 2021, on two mainstream and fringe Web communities. By analyzing 8B posts from Reddit and 206M posts from 4chan's /pol/, we investigate the origins, evolution, and content of Sinophobia. We find that, anti-Chinese content may be evoked by political events not directly related to China, e.g., the U.S. withdrawal from the Paris Agreement. And during the COVID-19 pandemic, daily usage of Sinophobic slurs has significantly increased even with the hate-speech ban policy. We also show that the semantic meaning of the words "China" and "Chinese" are shifting towards Sinophobic slurs with the rise of COVID-19 and remain the same in the pandemic period. We further use topic modeling to show the topics of Sinophobic discussion are pretty diverse and broad. We find that both Web communities share some common Sinophobic topics like ethnics, economics and commerce, weapons and military, foreign relations, etc. However, compared to 4chan's /pol/, more daily life-related topics including food, game, and stock are found in Reddit. Our finding also reveals that the topics related to COVID-19 and blaming the Chinese government are more prevalent in the pandemic period. To the best of our knowledge, this paper is the longest quantitative measurement of Sinophobia., Comment: To Appear in the 16th International Conference on Web and Social Media (ICWSM), 2022

Self-supervised learning is an emerging machine learning paradigm. Compared to supervised learning which leverages high-quality labeled datasets, self-supervised learning relies on unlabeled datasets to pre-train powerful encoders which can then be treated as feature extractors for various downstream tasks. The huge amount of data and computational resources consumption makes the encoders themselves become the valuable intellectual property of the model owner. Recent research has shown that the machine learning model's copyright is threatened by model stealing attacks, which aim to train a surrogate model to mimic the behavior of a given model. We empirically show that pre-trained encoders are highly vulnerable to model stealing attacks. However, most of the current efforts of copyright protection algorithms such as watermarking concentrate on classifiers. Meanwhile, the intrinsic challenges of pre-trained encoder's copyright protection remain largely unstudied. We fill the gap by proposing SSLGuard, the first watermarking scheme for pre-trained encoders. Given a clean pre-trained encoder, SSLGuard injects a watermark into it and outputs a watermarked version. The shadow training technique is also applied to preserve the watermark under potential model stealing attacks. Our extensive evaluation shows that SSLGuard is effective in watermark injection and verification, and it is robust against model stealing and other watermark removal attacks such as input noising, output perturbing, overwriting, model pruning, and fine-tuning., Comment: Accepted by CCS 2022

Self-supervised representation learning techniques have been developing rapidly to make full use of unlabeled images. They encode images into rich features that are oblivious to downstream tasks. Behind their revolutionary representation power, the requirements for dedicated model designs and a massive amount of computation resources expose image encoders to the risks of potential model stealing attacks - a cheap way to mimic the well-trained encoder performance while circumventing the demanding requirements. Yet conventional attacks only target supervised classifiers given their predicted labels and/or posteriors, which leaves the vulnerability of unsupervised encoders unexplored. In this paper, we first instantiate the conventional stealing attacks against encoders and demonstrate their severer vulnerability compared with downstream classifiers. To better leverage the rich representation of encoders, we further propose Cont-Steal, a contrastive-learning-based attack, and validate its improved stealing effectiveness in various experiment settings. As a takeaway, we appeal to our community's attention to the intellectual property protection of representation learning techniques, especially to the defenses against encoder stealing attacks like ours., Comment: Accepted by CVPR 2023

Many real-world data come in the form of graphs. Graph neural networks (GNNs), a new family of machine learning (ML) models, have been proposed to fully leverage graph data to build powerful applications. In particular, the inductive GNNs, which can generalize to unseen data, become mainstream in this direction. Machine learning models have shown great potential in various tasks and have been deployed in many real-world scenarios. To train a good model, a large amount of data as well as computational resources are needed, leading to valuable intellectual property. Previous research has shown that ML models are prone to model stealing attacks, which aim to steal the functionality of the target models. However, most of them focus on the models trained with images and texts. On the other hand, little attention has been paid to models trained with graph data, i.e., GNNs. In this paper, we fill the gap by proposing the first model stealing attacks against inductive GNNs. We systematically define the threat model and propose six attacks based on the adversary's background knowledge and the responses of the target models. Our evaluation on six benchmark datasets shows that the proposed model stealing attacks against GNNs achieve promising performance., Comment: To Appear in the 43rd IEEE Symposium on Security and Privacy, May 22-26, 2022

Many real-world data comes in the form of graphs, such as social networks and protein structure. To fully utilize the information contained in graph data, a new family of machine learning (ML) models, namely graph neural networks (GNNs), has been introduced. Previous studies have shown that machine learning models are vulnerable to privacy attacks. However, most of the current efforts concentrate on ML models trained on data from the Euclidean space, like images and texts. On the other hand, privacy risks stemming from GNNs remain largely unstudied. In this paper, we fill the gap by performing the first comprehensive analysis of node-level membership inference attacks against GNNs. We systematically define the threat models and propose three node-level membership inference attacks based on an adversary's background knowledge. Our evaluation on three GNN structures and four benchmark datasets shows that GNNs are vulnerable to node-level membership inference even when the adversary has minimal background knowledge. Besides, we show that graph density and feature similarity have a major impact on the attack's success. We further investigate two defense mechanisms and the empirical results indicate that these defenses can reduce the attack performance but with moderate utility loss.

Data is the key factor to drive the development of machine learning (ML) during the past decade. However, high-quality data, in particular labeled data, is often hard and expensive to collect. To leverage large-scale unlabeled data, self-supervised learning, represented by contrastive learning, is introduced. The objective of contrastive learning is to map different views derived from a training sample (e.g., through data augmentation) closer in their representation space, while different views derived from different samples more distant. In this way, a contrastive model learns to generate informative representations for data samples, which are then used to perform downstream ML tasks. Recent research has shown that machine learning models are vulnerable to various privacy attacks. However, most of the current efforts concentrate on models trained with supervised learning. Meanwhile, data samples' informative representations learned with contrastive learning may cause severe privacy risks as well. In this paper, we perform the first privacy analysis of contrastive learning through the lens of membership inference and attribute inference. Our experimental results show that contrastive models trained on image datasets are less vulnerable to membership inference attacks but more vulnerable to attribute inference attacks compared to supervised models. The former is due to the fact that contrastive models are less prone to overfitting, while the latter is caused by contrastive models' capability of representing data samples expressively. To remedy this situation, we propose the first privacy-preserving contrastive learning mechanism, Talos, relying on adversarial training. Empirical results show that Talos can successfully mitigate attribute inference risks for contrastive models while maintaining their membership privacy and model utility., Comment: To Appear in 2021 ACM SIGSAC Conference on Computer and Communications Security, November 2021

Inference attacks against Machine Learning (ML) models allow adversaries to learn sensitive information about training data, model parameters, etc. While researchers have studied, in depth, several kinds of attacks, they have done so in isolation. As a result, we lack a comprehensive picture of the risks caused by the attacks, e.g., the different scenarios they can be applied to, the common factors that influence their performance, the relationship among them, or the effectiveness of possible defenses. In this paper, we fill this gap by presenting a first-of-its-kind holistic risk assessment of different inference attacks against machine learning models. We concentrate on four attacks -- namely, membership inference, model inversion, attribute inference, and model stealing -- and establish a threat model taxonomy. Our extensive experimental evaluation, run on five model architectures and four image datasets, shows that the complexity of the training dataset plays an important role with respect to the attack's performance, while the effectiveness of model stealing and membership inference attacks are negatively correlated. We also show that defenses like DP-SGD and Knowledge Distillation can only mitigate some of the inference attacks. Our analysis relies on a modular re-usable software, ML-Doctor, which enables ML model owners to assess the risks of deploying their models, and equally serves as a benchmark tool for researchers and practitioners.

Graph data, such as chemical networks and social networks, may be deemed confidential/private because the data owner often spends lots of resources collecting the data or the data contains sensitive information, e.g., social relationships. Recently, neural networks were extended to graph data, which are known as graph neural networks (GNNs). Due to their superior performance, GNNs have many applications, such as healthcare analytics, recommender systems, and fraud detection. In this work, we propose the first attacks to steal a graph from the outputs of a GNN model that is trained on the graph. Specifically, given a black-box access to a GNN model, our attacks can infer whether there exists a link between any pair of nodes in the graph used to train the model. We call our attacks link stealing attacks. We propose a threat model to systematically characterize an adversary's background knowledge along three dimensions which in total leads to a comprehensive taxonomy of 8 different link stealing attacks. We propose multiple novel methods to realize these 8 attacks. Extensive experiments on 8 real-world datasets show that our attacks are effective at stealing links, e.g., AUC (area under the ROC curve) is above 0.95 in multiple cases. Our results indicate that the outputs of a GNN model reveal rich information about the structure of the graph used to train the model., Comment: To appear in the 30th Usenix Security Symposium, August 2021, Vancouver, B.C., Canada

Despite continuous efforts to build and update network infrastructure, mobile devices in developing regions continue to be constrained by limited bandwidth. Unfortunately, this coincides with a period of unprecedented growth in the size of mobile applications. Thus it is becoming prohibitively expensive for users in developing regions to download and update mobile apps critical to their economic and educational development. Unchecked, these trends can further contribute to a large and growing global digital divide. Our goal is to better understand the source of this rapid growth in mobile app code size, whether it is reflective of new functionality, and identify steps that can be taken to make existing mobile apps more friendly bandwidth constrained mobile networks. We hypothesize that much of this growth in mobile apps is due to poor resource/code management, and do not reflect proportional increases in functionality. Our hypothesis is partially validated by mini-programs, apps with extremely small footprints gaining popularity in Chinese mobile networks. Here, we use functionally equivalent pairs of mini-programs and Android apps to identify potential sources of "bloat," inefficient uses of code or resources that contribute to large package sizes. We analyze a large sample of popular Android apps and quantify instances of code and resource bloat. We develop techniques for automated code and resource trimming, and successfully validate them on a large set of Android apps. We hope our results will lead to continued efforts to streamline mobile apps, making them easier to access and maintain for users in developing regions., Comment: 12 pages, 8 figures

It is a widely accepted fact that state-sponsored Twitter accounts operated during the 2016 US presidential election spreading millions of tweets with misinformation and inflammatory political content. Whether these social media campaigns of the so-called "troll" accounts were able to manipulate public opinion is still in question. Here we aim to quantify the influence of troll accounts and the impact they had on Twitter by analyzing 152.5 million tweets from 9.9 million users, including 822 troll accounts. The data collected during the US election campaign, contain original troll tweets before they were deleted by Twitter. From these data, we constructed a very large interaction graph; a directed graph of 9.3 million nodes and 169.9 million edges. Recently, Twitter released datasets on the misinformation campaigns of 8,275 state-sponsored accounts linked to Russia, Iran and Venezuela as part of the investigation on the foreign interference in the 2016 US election. These data serve as ground-truth identifier of troll users in our dataset. Using graph analysis techniques we qualify the diffusion cascades of web and media context that have been shared by the troll accounts. We present strong evidence that authentic users were the source of the viral cascades. Although the trolls were participating in the viral cascades, they did not have a leading role in them and only four troll accounts were truly influential., Comment: With this version, we are correcting an error in the Acknowledgments regarding the research funding that supports this work. The correct one is the European Union's Horizon 2020 Research and Innovation program under the Cybersecurity CONCORDIA project (Grant Agreement No. 830927)