Federated unlearning (FU) allows a participating client in a federated learning (FL) system to remove its contribution from the trained global model, thereby enforcing the client’s ``right to be forgotten'' (RTBF). However, from the perspective of a client that does not request unlearning, the activation of the FU process may disrupt ongoing FL training and introduce additional computational and time overhead. In such cases, a client opposed to unlearning may be incentivized to retaliate against the unlearning client(s). In this work, we take the first step toward demonstrating the feasibility of such retaliatory behavior by exploiting the information leakage introduced during the FU process. Specifically, we propose a novel unlearning-induced membership inference attack (MIA) model, followed by a coarse-to-fine data generation method that enables an adversarial client to locally reconstruct the unlearned data. Building on this reconstruction, we introduce two targeted retaliatory attacks: (1) Anti-Unlearning Attack (AUA), which hinders the global model from successfully forgetting the data intended for removal, and (2) Discrimination-Unlearning Attack (DUA), which specifically degrades the global model’s performance on the unlearned data. Extensive experiments across a variety of FU methods and settings validate the effectiveness of the proposed retaliatory attack framework.

How to defend against attacks in Federated Partial Label Learning (FedPLL) is a brand new and challenging question in machine learning security due to stealthy and efficient attack behaviors of adversaries. In this paper, we systematically study this problem by developing an Adaptive Partial Label Attack (APLA) which subtly manipulates the candidate label set of the data sample. To defend against APLA, we develop the RobustFedPLL framework incorporating three modules: (1) in preliminary clustering, we implement a Gaussian Mixture Model (GMM) and a moving average mechanism to identify clients' confidence; (2) in representation contrasting, we develop a contrast-based algorithm to obtain clients' model feature representations; (3) in final clustering, we utilize mainstream clustering algorithms to finally distinguish adversaries. Experiments of RobustFedPLL and SOTA defense algorithms based on two datasets are conducted, demonstrating the superiority of RobustFedPLL under various experimental settings.

Algorithmic fairness, emphasizing that machine learning algorithms should not discriminate against specific demographic groups, has recently gained increasing attention in the context of federated learning. Existing algorithmic fairness algorithms in federated learning primarily focus on mitigating inherent biases within the training data or biases introduced by non-IID data distributions. However, these methods heavily rely on the assumption of trust and cooperation among participants in federated learning, leaving a notable gap in exploring the robustness of algorithmic fairness under adversarial environments. Therefore, our study advocates the exploration of the robustness of algorithmic fairness in federated learning. We first introduce four types of fairness attacks targeting the fairness of federated learning models. Then, we propose FairGuard, an innovative defense framework designed to counter these attacks. FairGuard employs a novel dual-inference algorithm that effectively identifies malicious clients during the global aggregation phase of federated learning. Our comprehensive experiments demonstrate that the introduced fairness attacks significantly compromise the global model's fairness and cannot be effectively mitigated by existing defense mechanisms, while FairGuard exhibits remarkable effectiveness in defending against these fairness attacks.

Federated unlearning (FU) algorithms offer participants in federated learning (FL) the "right to be forgotten'' for their individual data and its impact on a collaboratively trained model. Existing FU algorithms primarily focus on accelerating the retraining process and enhancing the utility of the retrained models following data removal requests. However, these approaches generally lack consideration for the robustness of FU algorithms in potential adversarial environments, where adversaries can craft malicious data removal requests to compromise the retrained model. In this work, we introduce a robust federated unlearning framework (robustFU) which notably enhances the resilience of FU algorithms against a wide range of adversarial attacks. In robustFU, we design a novel dynamic conflict sample compensation algorithm that dynamically reintroduces randomly generated samples with significant information gain to the participating clients during retraining. Additionally, robustFU employs an innovative global reweighting mechanism which adjusts the weight of each model update during the global aggregation, based on its degree of misalignment with the trained model prior to unlearning. Extensive experiments demonstrates the effectiveness and robustness of the proposed robustFU framework under adversarial environments. Furthermore, robustFU significantly accelerates the retraining process, achieving a 2.53× speed-up compared to the retrain from the scratch baseline.

IoT (Internet of Things) cameras have widely been deployed over the last few years. These cameras are often with limited hardware so that they can only capture noisy videos in low resolution. In this work, we propose the joint video denoising and super-resolution network for IoT cameras, which consists of the noise-robust moving-attention (NRMA) module and the noise-eliminated upsampling (NEU) module. In NRMA, we adopt a coarse-to-fine approach by first extracting the coarse flow and then refining through bi-directional feature propagation among adjacent frames. In NEU, we further utilize inner-frame features for noise-elimination and upsampling. Through this approach, we avoid the negative effects brought by applying denoising and super-resolution in tandem, and enhance the reconstruction of moving objects by the embedded attention layers in NRMA. We conduct our experiments on both synthetic datasets, which utilize existing data with additive white Gaussian noise (AWGN), and a realistic dataset captured using a pair of IoT and professional cameras. Our extensive experimental results demonstrate that our proposed method significantly reduces noise and enhances detail in both types of datasets. Notably, our approach outperforms the state-of-the-art benchmark (RealBasicVSR) by an average of 5.24 dB on the existing datasets (with noise level σ = 20) and by 0.95 dB on the realistic dataset in terms of PSNR.

Real-time video surveillance cameras have been widely deployed over the last few years. In case of incidents such as natural disasters, it provides vital guidance in real time to aid the rescue operations. However, the quality of the captured video is far from satisfactory due to the limited camera hardware and low network bandwidth. Noise is often observed especially at night and the resolution is low. To this end, we are motivated to retrieve the nighttime surveillance video through calibrated denoising and super-resolution. We only use the preceding and current frames, but not the future frames. Thus, we avoid the additional delay of waiting for future frames, which is not suitable for real-time video applications. We design the pipeline semantically beneficial for both denoising and super-resolution, and achieve high rendering quality, especially for real-world noise. Moreover, we propose a novel calibration method for collecting paired noisy and clean observations in the real world, which provides more effective training data. We conduct experiments using the real-world dataset collected under low-light conditions, and benchmark with state-of-the-art video denoising and super-resolution methods. Results show that our method achieves significant performance gain while introducing small delay compared with the benchmarks, suitable for real-time videos.