In medical image classification, data privacy constraints and the high cost of expert annotations pose significant challenges to building generalizable models. Federated semi-supervised learning (FSSL), which combines the privacy-preserving nature of federated learning with the label efficiency of semi-supervised learning, offers a promising direction. However, in real-world deployments, client data often exhibits highly non-independent and identically distributed (Non-IID) characteristics. This distributional heterogeneity undermines the reliability of pseudo-labels generated by global models, ultimately limiting model generalization. A key limitation of existing FSSL approaches lies in their reliance on a static labeled set fixed prior to training. Such strategies lack the ability to adaptively correct pseudo-label noise or address class imbalance throughout training, particularly under Non-IID settings. To address this, we propose FSSAL, a novel framework that introduces an active learning component into the FSSL pipeline. By continuously identifying informative and representative samples during training, our method adaptively refines the labeled set and enhances the model’s robustness to distribution shifts. FSSAL employs client-private models for pseudo-label generation to reduce global bias, applies a class-aware dynamic thresholding mechanism to ensure more reliable and balanced label selection, and incorporates a sample selection strategy guided by both feature diversity and model uncertainty. Extensive experiments on four public medical image classification datasets demonstrate that FSSAL consistently outperforms competitive FSSL methods in accuracy and F1-score, especially under highly Non-IID conditions, highlighting its robustness and practical potential.

Current approaches to memory in Large Language Models (LLMs) predominantly rely on static Retrieval-Augmented Generation (RAG), which often results in scattered retrieval and fails to capture the structural dependencies required for complex reasoning.For autonomous agents, these passive and flat architectures lack the cognitive organization necessary to model the dynamic and associative nature of longterm interaction.To address this, we propose Structured Episodic Event Memory (SEEM), a hierarchical framework that synergizes a graph memory layer for relational facts with a dynamic episodic memory layer for narrative progression.Grounded in cognitive frame theory, SEEM transforms interaction streams into structured Episodic Event Frames (EEFs) anchored by precise provenance pointers.Furthermore, we introduce an agentic associative fusion and Reverse Provenance Expansion (RPE) mechanism to reconstruct coherent narrative contexts from fragmented evidence.Experimental results on the LoCoMo and Long-MemEval benchmarks demonstrate that SEEM significantly outperforms baselines, enabling agents to maintain superior narrative coherence and logical consistency.

Abstract Federated learning (FL) is a type of distributed machine learning that enables multiple participants to collaboratively build machine learning models without transferring data outside their local devices, thereby ensuring data privacy and security. However, free-riding (FR) attacks pose significant threats by sending false, erroneous, or malicious model updates to the central server, attempting to extract private information from other devices during the federated learning process. This results in privacy leakage and reduced model accuracy. Traditional defenses measures against FR attacks typically employ auditing methods to identify malicious clients, but these methods are ineffective when multiple FR clients collude to inflate each other’s scores mutually. This paper proposes a novel defense method against collusion-based FR attacks. We first design a grouping mechanism based on gradient norm to group clients and then update the groups using an inter-client audit system. Finally, the correlation analysis of all groups is carried out to eliminate the attack group to ensure the security of the training process. This method defends against standard FR attacks and effectively detects attackers in collusion scenarios. Experimental results demonstrate that our method significantly improves the detection of malicious clients and enhances model accuracy by 10–20% compared to existing methods. Moreover, the proposed defense mechanism maintains its efficacy even in large-scale client environments, where more than 50% of the clients may be compromised by attackers.

This study investigates the state estimation problem of multi-agent systems under a homologous sensor attack. A distributed filter is proposed to achieve a minimum variance unbiased (MVU) estimation of system states and attacks in the presence of measurement noise. A gain matrix selection method for implementing the MVU estimation is also provided. The proposed filter can be used for positioning corrections affected by global positioning system (GPS) meaconing attacks. This study treats the positioning offset caused by GPS meaconing attacks as a zero-mean white random variable and verifies the validity of this hypothesis through experiments with real GPS signals. Moreover, this study comprehensively analyses the integration of the filters into practical systems. Finally, the effectiveness of the proposed results is verified using simulation examples. Note to Practitioners–This study introduces a filter that can achieve GPS positioning calibration under meaconing attacks. This filter treats the true positioning of the system as a state and the deviation caused by meaconing attacks as a homologous attack. The filter employs a collaborative filter to reconstruct the state, thereby enabling positioning calibration. Additionally, the study explores the relationship between the homology of meaconing attacks and the estimation error of filters. This result reveals that as the homology of attacks increases, the performance of filters also improves.

Abnormal working conditions often occur in modern industrial processes. Subsequently, if maintenance is not carried out in a timely manner, such abnormal working conditions can cause changes in system output variables, resulting in a decrease in product quality and even production accidents. This study proposes a self-healing control framework that extracts system input and output variable historical data under normal conditions. Then, based on simple correlation analyses and canonical correlation analyses, the correlation between historical input and output data is constructed, and the self-healing control law is calculated using the current batch of system output. The setpoint of the underlying controller and process variables are adjusted according to the control law to restore the system output variables to normal or near normal, thus ensuring the stability of product quality. This study demonstrates the convergence of system output variables and analyzes the computational complexity of the two algorithms. Finally, applied to water storage and continuous stirring reactor systems, the proposed self-healing control framework is validated to have a significant effect in restoring system output variables to normal or near normal, and presents significant safety and economic value in industrial production. Note to Practitioners—Abnormal conditions and fault conditions often occur in traditional complex industrial production processes, such as temperature, pressure fluctuations and pipeline blockages. When abnormal working conditions occur, certain key variables will fluctuate, which will have a huge impact on system safety and stable production. Therefore, this paper proposes a self-healing control system. The system can timely detect the occurrence of abnormal working conditions and make corresponding self-healing control decisions for abnormal working conditions. This system is compatible with several types of algorithms, and two different algorithms are designed in this article. Both algorithms can maintain the stability of key system output variables, and each has its own characteristics of adjustment and operation. In order to explore the application scope of the system, this paper carried out tests in various application scenarios and abnormal working conditions. Experimental results show that the proposed self-healing control system has significant self-healing effects in various industrial production processes.

In this study, a novel adaptive resilient control scheme is proposed to solve the prescribed-time stabilization problem of switched nonlinear systems under false data injection (FDI) attacks. A switched nonlinear state observer (SNSO) is designed to reconstruct unmeasurable system states. Additionally, a novel coordinate transformation and a Nussbaum function are introduced to address the challenges posed by FDI attacks on feedback control channels. The command filter is incorporated into the backstepping control framework, resulting in an adaptive resilient controller to ensure the stability and robustness of the system. Compared with the existing stability results of switched nonlinear systems, this study introduces an innovative prescribed-time scale transformation function into the SNSO and the adaptive resilient controller, thereby yielding a more relaxed criterion for user-defined settling time in the absence of prior information. The proposed piecewise switched adaptive laws reduce the conservatism of the designed controller in the presence of arbitrary switching. Finally, the feasibility of the proposed control theory is validated via simulations.

This study addresses the problem of the estimation of state when heterogeneous multiagent systems are affected by homologous unknown inputs (UIs). Homologous UIs refer to identical UIs affecting different agents. An improved semidistributed filter based on previous research is proposed. The improved filter uses neighbors’ information for UI estimation but not state estimation. A necessary and sufficient condition for the proposed filter to achieve minimum-variance unbiased estimation is presented and proven. Moreover, the asymptotic stability of the filter is analyzed. A sufficient condition of the asymptotic stability is presented and proven. The theoretical and numerical analyses indicate that the proposed filter has less communication pressure, fewer calculation requirements, and better estimation performance compared with the existing solutions.Note to Practitioners—In the industry, homologous unknown inputs (UIs) exist in many different systems. For example, the same ambient temperature affects the performance of every battery in a battery pack. Similarly, the same wind power can affect different aircrafts flying in the same region. Temperature and wind power can be considered the homologous UIs of a multiagent system. Estimation of homologous UIs is important because of the latter’s massive impact on the system. In this study, data transmission delay and packet loss are ignored. Hence, the study is limited to low-rate systems. Moreover, nonlinear filters must be studied further in future work.

Secure state estimation (SSE) is a problem to defense false-data injection attacks. This study designs an online distributed SSE of heterogeneous multiagent systems under homologous attack. A triple-loop observer is proposed to estimate the state and attack signal simultaneously. The inner loop keeps the estimations of the attack signal the same using average consensus. The middle loop adjusts the estimations via residual information. The outer loop runs when system measurements change. A sufficient condition that estimations asymptotically converge to the real value is obtained and proved. Finally, the proposed observer has been tested on the global positioning system.