The contemporary research paradigm in remote sensing hyperspectral monitoring increasingly relies on multisatellite and multiplatform Earth observation. While the traditional hyperspectral research framework based on single-image processing (SIP) has facilitated the application of idealized scenarios and the development of standardized evaluation benchmarks, it inherently constrains the model's ability to generalize feature representations across varying spatial and temporal domains. As hyperspectral data applications grow in complexity and data requirements, the limitations of SIP in addressing the demands of modern remote sensing tasks become increasingly apparent. To overcome these research limitations, we utilize the decentralized nature and data security features of federated learning to propose a cross-image hyperspectral image (HSI) federated learning approach for classification tasks. We first develop a client-oriented self-guided knowledge-enhanced personalized learning method that enhances the personalization of the local learning process by leveraging relevant features from other clients, thereby improving the learning efficiency of each client. To address the issue of "bias" in global knowledge caused by uneven data distribution across the federated learning process, we introduce a multiscale semantic aligned dynamic aggregation method to ensure fairness in integrating global knowledge. To our knowledge, this article is the first to explore the joint learning of HSI classification using federated learning. Accordingly, we have constructed open-set and closed-set datasets tailored to this task and have demonstrated the effectiveness of our method on these datasets. The code is available at: https://github.com/Gallipaxi/FedHIC.

6G plans to deeply integrate Artificial Intelligence (AI) to support large-scale device deployment and optimize network management by predicting the UE-connected gNB. However, user prediction in 6 G mobile communication networks faces three major challenges, temporal complexity, spatial constraints, and latency sensitivity. This paper introduces a high-accuracy, high-efficient, and low-latency user prediction model, called RelNet, designed to address these challenges. RelNet consists of three modules. The UE Trajectory Patching (UTP) module first reduces computational complexity by transmitting UE trajectory data into patches. The Temporal Feature Extraction (TFE) module then captures both short-term fluctuations and long-term dependencies in massive UE trajectories. Finally, the Intelligent Positioning Refinement (IPR) module enables adaptive optimization and precise gNB positioning. Furthermore, this paper uses User Prediction Assisted Paging (UPAP) to validate the practicality of RelNet in 6 G, reducing signaling overhead by predicting the next-connected gNB and generate the secondary paging area (SPA). Experimental results demonstrate that RelNet achieves 69.98 % prediction accuracy, outperforming mainstream models. Moreover, the UPAP scheme effectively reduces signaling overhead by 69.72 % on a real-world 6 G dataset, confirming the feasibility of RelNet in 6 G networks.

Tiny crack detection plays a vital role in ensuring the safety of critical industrial components and infrastructure, providing early intervention to prevent crack propagation and mitigate potential damage. Although machine vision-based defect detection has been greatly utilized in safety inspection and maintenance, tiny crack detection remains a big issue due to the low visibility and weak features with large background noise. To address this challenge, this paper presents an end-to-end trainable hierarchical constrained neural network for tiny crack detection. Firstly, we present a hierarchical loss constraint Filter (HLCF) module based on a region-level attention mechanism to capture the holistic vein structure of cracks and enhance the faint feature extraction of tiny cracks. In order to balance local high contrast and holistic crack features, a feature fusion loss constraint is designed to reduce noise interference and refine the boundary details of cracks by learning different receptive fields in each layer. Besides that, we created a dataset consisting of tiny crack samples collected through image processing and manual labeling from industrial production, named TinyCrack. The proposed HLCF model is evaluated on TinyCrack and seven public databases. The experimental indices show that the model achieves Precision over 0.62%, Recall over 0.12%, F-score over 0.53%, and IOU over 1.12% in TinyCrack, demonstrating the best accuracy compared with other public crack detection models. The results indicate that the HLCF detection model has better performance in identifying low-visibility tiny cracks.

The problem of designing explicit bent and plateaued functions has been researched for several decades. However, finding new bent functions outside the well-known completed Maiorana-McFarland class$\mathcal {M}^{\#}$is still a challenge. Plateaued functions have been characterized in many different ways, but there is no general and rigorous mathematical method to generate them directly, except for the ones in the spirit of the well-known Maiorana-McFarland constructions or those obtained through adaptations of the secondary constructions of bent functions. Jeong and Lee recently made significant advances regarding algorithms for constructing balanced plateaued functions with maximal algebraic degrees in [IEEE Trans. Inf. Theory, 70(2), 1408-1421, 2024]. Due to the gap between our significant interest in the notion of plateaued functions and the knowledge we have on it, our motivation is to bring further results on the constructions of plateaued functions that allow us to understand their structure better. This article creates a framework of new generic constructions of bent and plateaued functions by studying Boolean functions of the form$h(x)=f(x)+F(f_{1}(x),\ldots, f_{r}(x))$, where$f_{i}(x)=f(x)+f(x+\mu _{i})$for each$1\leq i\leq r$. We firstly prove that h and f have the same extended Walsh-Hadamard spectrum if$D_{\mu _{i}}D_{\mu _{j}}f=0$for any$1\leq i\lt j\leq r$. This result extends a previous construction of bent functions to any Boolean functions. The strength of such a result is that it allows us to obtain several plateaued functions of high algebraic degrees from known ones with low algebraic degrees, which was a significant and challenging problem raised in the literature. Such a result is a real challenge and breaks a deadlock since no mathematical method allows the general constructions of plateaued functions. We next give an extended affine equivalent form of the function h, which provides us with another compelling perspective to design new bent functions (including those which are outside$\mathcal {M}^{\#}$from certain known ones inside$\mathcal {M}^{\#}$) and plateaued functions. Finally, we present four generic constructions of bent functions outside$\mathcal {M}^{\#}$from generalized Maiorana-McFarland functions.

Recently, much progress has been made to construct minimal linear codes due to their preference in secret sharing schemes and secure two-party computation. In this paper, we put forward a new method to construct minimal linear codes by using vectorial Boolean functions. Firstly, we give a necessary and sufficient condition for a generic class of linear codes from vectorial Boolean functions to be minimal. Based on that, we derive some new three-weight minimal linear codes and determine their weight distributions. Secondly, by studying deeply the construction of linear codes in this paper, we find a necessary and sufficient condition of the linear codes to be minimal and to be violated the AB condition. As a result, we get three infinite families of minimal linear codes violating the AB condition. To the best of our knowledge, this is the first time that minimal liner codes are constructed from vectorial Boolean functions. Compared the parameters with other known ones, in general the minimal liner codes obtained in this paper have higher dimensions.

Twitter bot detection has become an increasingly important task to combat misinformation, facilitate social media moderation, and preserve the integrity of the online discourse. State-of-the-art bot detection methods generally leverage the graph structure of the Twitter network, and they exhibit promising performance when confronting novel Twitter bots that traditional methods fail to detect. However, very few of the existing Twitter bot detection datasets are graph-based, and even these few graph-based datasets suffer from limited dataset scale, incomplete graph structure, as well as low annotation quality. In fact, the lack of a large-scale graph-based Twitter bot detection benchmark that addresses these issues has seriously hindered the development and evaluation of novel graph-based bot detection approaches. In this paper, we propose TwiBot-22, a comprehensive graph-based Twitter bot detection benchmark that presents the largest dataset to date, provides diversified entities and relations on the Twitter network, and has considerably better annotation quality than existing datasets. In addition, we re-implement 35 representative Twitter bot detection baselines and evaluate them on 9 datasets, including TwiBot-22, to promote a fair comparison of model performance and a holistic understanding of research progress. To facilitate further research, we consolidate all implemented codes and datasets into the TwiBot-22 evaluation framework, where researchers could consistently evaluate new models and datasets. The TwiBot-22 Twitter bot detection benchmark and evaluation framework are publicly available at \url{https://twibot22.github.io/}.

This article is devoted to Boolean and vectorial bent functions and their duals. Our ultimate objective is to increase such functions’ corpus by designing new ones covering many previous bent functions’ constructions. To this end, we provide several new infinite families of bent functions, including idempotent bent functions of any algebraic degree, bent functions in univariate trace form, and self-dual bent functions. Those bent functions are of great theoretical and practical interest because of their special structures and relationship with self-dual codes. In particular, many well-known bent functions are special cases of our bent functions. Moreover, we extend our results to vectorial bent functions and obtain three new infinite classes of vectorial bent functions of any possible degree by determining the explicit duals of three classes of well-known bent functions.

Let$n=2m$. In 2020, Budaghyan, Helleseth and Kaleyski [IEEE TIT 66(11): 7081-7087, 2020] considered a family of quadrinomials over$\mathbb {F}_{2^{n}}$of the form$x^{3}+a(x^{2^{s}+1})^{2^{k}}+bx^{3\cdot 2^{m}}+c(x^{2^{s+m}+2^{m}})^{2^{k}}$. They showed that two infinite classes of almost perfect nonlinear (APN) functions belong to this family when$\gcd (6,m)=1$. We observe that these two infinite classes of APN quadrinomials and the infinite class of APN polynomials from the Budaghyan-Carlet family belong to a more general family of polynomials over$\mathbb {F}_{2^{n}} $with the form$f(x)=a{\mathrm{ Tr}}^{n}_{m}(F(x))+a^{2^{m}}{\mathrm{ Tr}}^{n}_{m}(G(x))$, where$a \in \mathbb {F}_{2^{n}}\backslash \mathbb {F}_{2^{m}} $, and both$F$and$G$are quadratic functions over$\mathbb {F}_{2^{n}}$. We characterize when$f(x) $is APN. With the help of our characterization, letting$F(x)=bx^{2^{i}+1} $and$G(x)=cx^{2^{s}+1}$with$b, c\in \mathbb {F}_{2^{n}} $, we obtain an infinite family of APN functions of the form$f(x) $when${\mathrm{ gcd}}(2,m)=1 $and verify that for$n=10 $two APN instances from this infinite family are CCZ-inequivalent to each other, and to any APN function over$\mathbb {F}_{2^{10}} $from the previously known infinite families.

Abstract In 2012, Carlet et al. developed two secondary constructions of bent functions (Advances in Mathematics of Communications, 6: 305‐314) and proposed some applications for their constructions. However, the duals of bent functions in their constructions were not presented. In order to find more general applications to these constructions and obtain new classes of bent functions, an open problem was proposed by Carlet in 2014. Hence, in this study, a class of vectorial bent functions for answering that open problem, which also addresses another open problem on vectorial bent functions proposed by Mesnager in 2014, is constructed. In addition, a new secondary construction of bent functions that generalises one of Carlet et al.'s constructions in 2012 is presented. Based on that, two new classes of bent functions were obtained and their duals were presented explicitly. In particular, some self‐dual bent functions are constructed. Moreover, it can be proved that our bent functions can be EA‐inequivalent to those constructed by Carlet et al. in 2012.