The rapid expansion of electric vehicles (EVs) charging stations underscores the urgent need for robust anomaly detection systems capable of identifying potential malfunctions while preserving data privacy. Federated Learning (FL) has emerged as a promising solution, enabling collaborative model training without requiring raw data sharing. However, applying conventional FL approaches to charging station networks presents significant challenges, including non-independent and identically distributed (non-IID) data and gradient conflicts among clients. To address these challenges, we introduce FedPareto, a novel Pareto-optimal FL framework designed to manage gradient conflicts and counter malicious attacks in charging station anomaly detection. FedPareto features a gradient conflict-aware aggregation method, which adaptively adjusts client weights based on cosine similarity between gradients, and a gradient magnitude reshaping strategy to enhance model convergence. Theoretical analysis demonstrates that FedPareto achieves a convergence rate of O($\frac{1}{T}$) and attains Pareto-optimal solutions under standard smoothness and convexity assumptions. Extensive experiments on real-world charging station datasets validate FedPareto’s effectiveness. It outperforms state-of-the-art methods, exhibits better robustness against gradient-based attacks, and ensures equitable performance distribution across clients. These results highlight FedPareto’s potential as a reliable and scalable solution for anomaly detection in EV charging station networks.

federated graph learning (FGL) has emerged as a promising paradigm for privacy-preserving collaborative learning in Social Internet of Things (SIoT), where nodes form complex interconnected networks. Existing FGL approaches face significant challenges including model degradation in handling nonindependent and identically distributed (non-IID) data and maintaining model performance across heterogeneous nodes. This article proposes framework via layer-wised clustering (FedLWC), a novel layer-wised clustering framework inspired by evolutionary processes is proposed to enhance the effectiveness of FGL. FedLWC designs three key aspects: 1) a fisher information matrix-based layer selection mechanism that identifies and evaluates critical model layers, which can reduce parameter redundancy; 2) a layer intersection clustering algorithm that preserves common key layers while accommodating local features; and 3) an adaptive layer merge strategy that effectively combines global shared layers with clustered key layers. To make sure that the proposed approach is rigorous, we conduct theoretical convergence analysis for the proposed framework under non-IID conditions. Extensive experiments on multiple benchmark graph datasets demonstrate FedLWC’s performance, achieving an average accuracy improvement of 7.01% compared to state-of-the-art federated learning methods.

Room-temperature superconductors (esp. LK-99 in the recent) have attracted extensive academic attention in recent years, both in academic circles and among the general public. This topic has spread through a number of social media channels, a plethora of contradiction in information has emerged within social networks. There arises the question on how to analyze the consistency and controversy of such scientific knowledge in the dissemination process, and how this process impact on public cognition on the scientific knowledge. In this article, taking room-temperature superconductor as example, we first designed a large language model based factual consistency detection approach to analyze the consistency between research papers and media reports. Then the consistency between media reports and comments is analyzed, by proposing a novel quantification method for media agenda-setting capability, which evaluates the agenda-setting capability of media based on emotional and positional consistencies. The results indicate that two significant deviations occur when room-temperature superconductor knowledge is spread from specialized fields to the public through the various media. One deviation is due to the specialized nature of room-temperature superconductor knowledge, leading to discrepancies between reported content and factual information in research papers. The other deviation is caused by conflicting knowledge, resulting in disparities between media reports and public perception.

Millions of patients suffer from rare diseases around the world. However, the samples of rare diseases are much smaller than those of common diseases. Hospitals are usually reluctant to share patient information for data fusion due to the sensitivity of medical data. These challenges make it difficult for traditional AI models to extract rare disease features for disease prediction. In this paper, we propose a Dynamic Federated Meta-Learning (DFML) approach to improve rare disease prediction. We design an Inaccuracy-Focused Meta-Learning (IFML) approach that dynamically adjusts the attention to different tasks according to the accuracy of base learners. Additionally, a dynamic weight-based fusion strategy is proposed to further improve federated learning, which dynamically selects clients based on the accuracy of each local model. Experiments on two public datasets show that our approach outperforms the original federated meta-learning algorithm in accuracy and speed with as few as five shots. The average prediction accuracy of the proposed model is improved by 13.28% compared with each hospital's local model.

The dynamics of public opinion on social media affects people’s feeling and minds about international affairs and leads to the reconstruction of societal states for international conflicts. In this article, we analyze the topics’ evolution on social media during the Pelosi visit. Such kind of analysis should help the related departments sense and beware the situation effectively and efficiently, and may provide technical supports for proper policy making and responses. To facilitate this purpose, a new method is proposed and an abbreviated large-graph clustering (ALGC) algorithm has been designed to generate documents and topic representation for alleviating the overhead of high computational complexity of large graphs by reducing the dimensionality of the attention matrix and adjacency matrix. The evolution pattern of topics is also analyzed in and between different time periods. Experiment results show that the proposed method performs well, achieving a high clustering accuracy with lower computational cost. The dataset used in this article is also released for public analysis.

Knowledge question and answering (Q&A) is widely used. However, most existing semantic parsing methods in Q&A usually use cascading, which can incur error accumulation. In addition, using only one institution’s Q&A data definitely will limit the Q&A performance, while data privacy prevents sharing between institutions. This article proposes a knowledge graph-based reinforcement federated learning (KGRFL)-based Q&A approach to address these challenges. We design an end-to-end multitask semantic parsing model [MSP-bidirectional and auto-regressive transformers (BART)] that identifies question categories while converting questions into SPARQL statements to improve semantic parsing. Meanwhile, a reinforcement learning (RL)-based model fusion strategy is proposed to improve the effectiveness of federated learning, which enables multi-institution joint modeling and data privacy protection using cross-domain knowledge. In particular, it also reduces the negative impact of low-quality clients on the global model. Furthermore, a prompt learning-based entity disambiguation method is proposed to address the semantic ambiguity problem because of joint modeling. The experiments show that the proposed method performs well on different datasets. The Q&A results of the proposed approach outperform the approach of using only a single institution. Experiments also demonstrate that the proposed approach is resilient to security attacks, which is required for real applications.

Examining the perceptual differences between China and the United States can facilitate a better understanding of their opinions and perspectives, helping to promote peaceful interactions among the two nations as well as the world. This study presents a data-driven approach to measure cognitive differences, investigating these differences from topical and sentimental angles regarding the unmanned balloon event that had been shot down by U.S. warplanes. We also explore the cognitive differences between news media and the followers, and the evolution of topics over time. Our findings reveal those discussions about “balloons” on social media in China and the United States display certain differences in terms of sentiment. In addition, we assess the impact of this event on U.S.–China relationship, particularly in trade. To evaluate the analytical capabilities of the popular ChatGPT model, we use this event as a case study to demonstrate that ChatGPT-like models may have limited capabilities for such kind of specialized analysis. The dataset utilized here is made available for public usage for further investigation on public opinion dynamics for similar events.

Based on an investigation of online public opinion on the Nahel Merzouk protests in France, an approach for analyzing and predicting public opinion on protests based on large language model (LLM) is proposed, revealing the impact of emerging social media on the protests. We demonstrate that protests generate public opinion on social media with some lag, but that comment sentiment and expression are consistent with protest trends. As the protests unfolded, we analyzed the evolution of public sentiment. We constructed the prompt based on historical data to predict the protests using the p-tuning and Lora approach to fine-tune LLM. In addition, we discuss how to use blockchain technology to optimize distributed, self-organizing protests and reduce the potential for disinformation and violent conflict.

The intensity and scale of the opinion fightings in cyberspace on the Russia–Ukraine war (RUW) have opened a new chapter in the history of world warfare. This is a magnificent demonstration of social cognitive war fighting with cyber-physical-social systems (CPSS) that would impact our humankind significantly now and for a long time to come, not just on our understanding of wars, but every aspect of our life. Therefore, it is worth of studying the opinion dynamics of the RUW in the cyberspace. This article will start this direction with an analysis of the evolutionary dynamics of the public opinion fighting, only using Chinese Weibo texts as a case study due to the time constraint. It first clusters the Weibo texts into four categories with unsupervised learning method using Latent Dirichlet Allocation and then collects opinions by extracting keywords. Meanwhile, an opinion adversarial evolution algorithm is proposed to dynamically model the dominance degree of an opinion in the evolutionary processes. We release a dataset of Chinese Weibo associated with RUW. The proposed approach of modeling and analyzing data-driven public opinion dynamics provides a new way for accessing opinion warfare in CPSS.