Fengyuan Yu

Traffic detection systems based on machine learning have been proposed to defend against cybersecurity threats, such as intrusion attacks and malware. However, they did not take the impact of network-induced phenomena into consideration, such as packet loss, retransmission, and out-of-order. These phenomena will introduce additional misclassifications in the real world. In this paper, we present \sf ERNN, a robust and end-to-end RNN model that is specially designed against network-induced phenomena. As its core, \sf ERNN is designed with a novel gating unit named as session gate that includes: (i) four types of actions to simulate common network-induced phenomena during model training; and (ii) the Mealy machine to update states of session gate that adjusts the probability distribution of network-induced phenomena. Taken together, \sf ERNN advances state-of-the-art by realizing the model robustness for network-induced phenomena in an error-resilient manner. We implement \sf ERNN and evaluate it extensively on both intrusion detection and malware detection systems. By practical evaluation with dynamic bandwidth utilization and different network topologies, we demonstrate that \sf ERNN can still identify 98.63% of encrypted intrusion traffic when facing about 16% abnormal packet sequences on a 10 Gbps dataplane. Similarly, \sf ERNN can still robustly identify more than 97% of the encrypted malware traffic in multi-user concurrency scenarios. \sf ERNN can realize \sim4% accuracy more than SOTA methods. Based on the Integrated Gradients method, we interpret the gating mechanism can reduce the dependencies on local packets (termed dependency dispersion). Moreover, we demonstrate that \sf ERNN possesses superior stability and scalability in terms of parameter settings and feature selection.

Federated learning achieves effective performance in modeling decentralized data. In practice client data are not well-labeled which makes it potential for federated unsupervised learning (FUSL) with non-IID data. However the performance of existing FUSL methods suffers from insufficient representations i.e. (1) representation collapse entanglement among local and global models and (2) inconsistent representation spaces among local models. The former indicates that representation collapse in local model will subsequently impact the global model and other local models. The latter means that clients model data representation with inconsistent parameters due to the deficiency of supervision signals. In this work we propose FedU2 which enhances generating uniform and unified representation in FUSL with non-IID data. Specifically FedU2 consists of flexible uniform regularizer (FUR) and efficient unified aggregator (EUA). FUR in each client avoids representation collapse via dispersing samples uniformly and EUA in server promotes unified representation by constraining consistent client model updating. To extensively validate the performance of FedU2 we conduct both cross-device and cross-silo evaluation experiments on two benchmark datasets i.e. CIFAR10 and CIFAR100.

Federated learning (FL) is a promising machine learning paradigm that collaborates with client models to capture global knowledge. However, deploying FL models in real-world scenarios remains unreliable due to the coexistence of in-distribution data and unexpected out-of-distribution (OOD) data, such as covariate-shift and semantic-shift data. Current FL researches typically address either covariate-shift data through OOD generalization or semantic-shift data via OOD detection, overlooking the simultaneous occurrence of various OOD shifts. In this work, we propose FOOGD, a method that estimates the probability density of each client and obtains reliable global distribution as guidance for the subsequent FL process. Firstly, SM3D in FOOGD estimates score model for arbitrary distributions without prior constraints, and detects semantic-shift data powerfully. Then SAG in FOOGD provides invariant yet diverse knowledge for both local covariate-shift generalization and client performance generalization. In empirical validations, FOOGD significantly enjoys three main advantages: (1) reliably estimating non-normalized decentralized distributions, (2) detecting semantic shift data via score values, and (3) generalizing to covariate-shift data by regularizing feature extractor. The prejoct is open in https://github.com/XeniaLLL/FOOGD-main.git.

With the rapid advancement of generative models, associated privacy concerns have attracted growing attention. To address this, researchers have begun adapting machine unlearning techniques from traditional classification models to generative settings. Although notable progress has been made in this area, a unified framework for systematically organizing and integrating existing work is still lacking. The substantial differences among current studies in terms of unlearning objectives and evaluation protocols hinder the objective and fair comparison of various approaches. While some studies focus on specific types of generative models, they often overlook the commonalities and systematic characteristics inherent in Generative Model Unlearning (GenMU). To bridge this gap, we provide a comprehensive review of current research on GenMU and propose a unified analytical framework for categorizing unlearning objectives, methodological strategies, and evaluation metrics. In addition, we explore the connections between GenMU and related techniques, including model editing, reinforcement learning from human feedback, and controllable generation. We further highlight the potential practical value of unlearning techniques in real-world applications. Finally, we identify key challenges and outline future research directions aimed at laying a solid foundation for further advancements in this field. We consistently maintain the related open-source materials at this https URL.