In this paper, we attempt to attack deep forgery detection systems rigorously to uncover their shortcomings. Unlike existing methods that solely attack fake regions to increase their authenticity, the proposed identifies decisive regions to decision and creates a dual mask to carry out different attacks. The dual mask comprises two components: a positive mask, which identifies real regions for targeted attacks, and a negative mask, pinpointing fake regions for disruption. Perturbations applied to the positive mask aim to bolster confidence in real decisions, while those on the negative mask aim to diminish confidence in fake decisions, thereby inducing the detection system to make erroneous judgments. Additionally, to enhance the visual quality of the manipulated samples, we introduce a hybrid constraint optimization, which carefully balances the advantages and disadvantages of spatial perturbations and frequency-domain adversarial samples, facilitating the learning of perturbations that maintain high attack efficacy while preserving visual fidelity. These two modules are integrated into a novel GAN-based attacking network, HPDM, to generate adversarial samples and achieve improvements in attack transferability and imperceptibility. We evaluated our method through experiments on FFHQ and FF++. The results demonstrate that our approach consistently outperforms existing methods in terms of average attack efficacy.

Voxel-level annotation has always been a burden of training medical image segmentation models. This paper investigates an interesting problem that finds the host organ of a lesion without actually labeling the organ. To remedy the missing annotation, we construct a graph using an off-the-shelf registration algorithm, on which lesion labels over the training set are accumulated to obtain the pseudo organ for each case. These pseudo labels are used to train a deep network, whose predictions determine the affinity of each lesion on the registration graph. We iteratively update the pseudo labels with the affinity until the training convergence. Our method is evaluated on the MSD Liver and KiTS datasets, without seeing any organ annotation, we achieve the test Dice score of 93% for liver and 92% for kidney, and boosts the accuracy of tumor segmentation to a considerable degree, $3%$, which even surpasses the model trained with ground-truth of both organ and tumor.

Reinforcement Learning (RL) has been applied to solve decision-making problems in computer network designs, especially in TCP congestion control. As RL-based congestion control methods enable powerful learning abilities, it achieves competitive performance and adaptiveness advantages over the traditional methods. However, RL-based systems suffer from adversarial attacks that generate perturbations to significantly degrade the performance. In this paper, we conduct a comprehensive study of adversarial attacks against RL-based congestion control systems. Unlike the state-of-the-art adversarial attacks on images where an attacker can easily obtain the input states to introduce perturbations, the attacker cannot directly obtain the input states in congestion control settings that are only available to the agents. It is challenging to add effective perturbations without knowing the input states for RL-based congestion control models. To solve the challenge, we develop an adversarial attack to estimate states of the target agent, craft adversarial perturbations, and apply the generated perturbations in an automated fashion. We evaluate how our adversarial attack affects the target agent’s decision-making process. Our experiments illustrate that our attack can effectively reduce about 50% average throughput while increasing more than 36x latency and 45% packet loss rate.

Unsupervised pretraining is of great significance for visual representation. Especially, contrastive learning has achieved great success recently, but existing approaches mostly ignored spatial information which is often crucial for visual representation. Strong semantic embedding has an inherent advantage for classification, but dense prediction tasks require more spatial and low-level representation. This paper presentsheterogeneous contrastive learning(HCL), an effective approach that adds spatial information to the encoding stage to alleviate the learning inconsistency between the contrastive objective and strong data augmentation operations. We demonstrate the effectiveness of HCL by showing that (i) it achieves higher accuracy in instance discrimination, (ii) it surpasses existing pre-training methods in a series of downstream tasks (iii) and it shrinks the pre-training costs by half for almost 800 GPU-hours. More importantly, we show that our approach achieves higher efficiency in visual representations, and thus delivers a key message to inspire the future research of self-supervised visual representation learning.

Privacy-centric mobile social network (PC-MSN), which allows users to build intimate and private social circles, is an increasingly popular type of online social networks (OSNs). Because of strict usage policy enforced by PC-MSNs (such as restricted account and content access), malicious accounts (or users) have to act like normal accounts to accumulate credentials before committing malicious activities. Therefore, analysis merely relying on static account profile information or social graphs is ineffective to detect such growing-up accounts. Besides, existing behavior-based malicious account detection methods fail to effectively detect growing-up accounts who pretend to be benign and have similar behaviors to benign users during the growing-up stage.

Semi-supervised learning is a useful tool for image segmentation, mainly due to its ability in extracting knowledge from unlabeled data to assist learning from labeled data. This paper focuses on a popular pipeline known as self-learning, where we point out a weakness named lazy mimicking that refers to the inertia that a model retains the prediction from itself and thus resists updates. To alleviate this issue, we propose the Asynchronous Teacher-Student Optimization (ATSO) algorithm that (i) breaks up continual learning from teacher to student and (ii) partitions the unlabeled training data into two subsets and alternately uses one subset to fine-tune the model which updates the labels on the other. We show the ability of ATSO on medical and natural image segmentation. In both scenarios, our method reports competitive performance, on par with the state-of-the-arts, in either using partial labeled data in the same dataset or transferring the trained model to an unlabeled dataset.

Online social networks (OSNs) are plagued by fake accounts. Existing fake account detection methods either require a manually labeled training set, which is time-consuming and costly, or rely on rich information of OSN accounts, e.g., content and behaviors, which incurs significant delay in detecting fake accounts. In this work, we propose UFA (Unveiling Fake Accounts) to detect fake accounts immediately after they are registered in an unsupervised fashion. First, through a measurement study on the registration patterns on a real-world registration dataset, we observe that fake accounts tend to cluster on outlier registration patterns, e.g., IP and phone numbers. Then, we design an unsupervised learning algorithm to learn weights for all registration accounts and their features that reveal outlier registration patterns. Next, we construct a registration graph to capture the correlation between registration accounts, and utilize a community detection method to detect fake accounts via analyzing the registration graph structure. We evaluate UFA using real-world WeChat datasets. Our results demonstrate that UFA achieves a precision 94% with a recall ~80%, while a supervised variant requires 600K manual labels to obtain the comparable performance. Moreover, UFA has been deployed by WeChat to detect fake accounts for more than one year. UFA detects 500K fake accounts per day with a precision ~93% on average, via manual verification by the WeChat security team.