Cyberswarm: a novel swarm intelligence algorithm inspired by cyber community dynamics

Received: 24 February 2025 / Accepted: 23 September 2025 / Published online: 4 November 2025 © The Author(s) 2025 Extended author information available on the last page of the article Cyberswarm: a novel swarm intelligence algorithm inspired by cyber community dynamics Abdelsadeq Elfergany1 · Ammar Adl1 · Mohammed Kayed1 Artificial Intelligence Review (2025) 58:406 https://doi.org/10.1007/s10462-025-11406-4 Abstract  Recommendation systems face challenges in dynamically adapting to evolving user preferences and interactions within complex social networks. Traditional approaches often fail to account for the intricate interactions within cyber-social systems and lack the flexibility to generalize across diverse domains, highlighting the need for more adaptive and versatile solutions. In this work, we introduce a general-purpose swarm intelligence algo­ rithm for recommendation systems, designed to adapt seamlessly to varying applications. It was inspired by social psychology principles. The framework models user preferences and community influences within a dynamic hypergraph structure. It leverages centrality-based feature extraction and Node2Vec embeddings. Preference evolution is guided by message- passing mechanisms and hierarchical graph modeling, enabling real-time adaptation to changing behaviors. Experimental evaluations demonstrated the algorithm’s superior per­ formance in various recommendation tasks, including social networks and content discov­ ery. Key metrics such as Hit Rate (HR), Mean Reciprocal Rank (MRR), and Normalized Discounted Cumulative Gain (NDCG) consistently outperformed baseline methods across multiple datasets. The model’s adaptability to dynamic environments allowed for contextu­ ally relevant and precise recommendations. The proposed algorithm represents an advance­ ment in recommendation systems by bridging individual preferences and community influences. Its general-purpose design enables applications in diverse domains, including social graphs, personalized learning, and medical graphs. This work highlights the potential of integrating swarm intelligence with network dynamics to address complex optimization challenges in recommendation systems. Graphic abstract  1 3 A. Elfergany et al. Keywords  CyberSwarm algorithm (CyS) · Swarm intelligence · Cyber community · Recommendation systems · GNN · Centrality measures 1  Introduction In this paper, we present our CyberSwarm Algorithm (CyS). It is a general-purpose swarm intelligence algorithm inspired by the interactive dynamics of social networks. It models each entity as a node within a graph, where edges represent connections like shared inter­ ests or exchanges of information, facilitating dynamic, context-sensitive interactions. The proposed model is guided by principles from social psychology, incorporating aspects of attitude and belief evolution. Drawing on theories such as Social Judgment and Consistency Theory, each node’s preferences evolve as a function of its interactions. Starting from an ini­ tial preference vector Pi(0), each node iteratively adjusts its preferences toward alignment with its social context. This adaptive process is encapsulated in a time-dependent preference vector Pi(t + 1), which integrates influence from neighboring nodes. The result is a flexible recommendation framework that adapts well to various applications. Our proposed framework incorporates a hypergraph (H) to enhance its performance further. Each node’s feature vector is enriched by centrality measures (Mcentral(H)) and Node2Vec embeddings. These features play a crucial role in capturing evolving preferences in real time. A central aspect of the proposed algorithm is the Dynamic Collaborative Swarm 1 3 406  Page 2 of 49 Cyberswarm: a novel swarm intelligence algorithm inspired by cyber… Equilibrium (DCSE) theorem. The DCSE theorem explains how equilibrium is reached in CyS, balancing both direct interactions and evolving social factors. This equilibrium is cru­ cial for maintaining stability and adaptability in complex environments, allowing the system to provide relevant recommendations even as the network evolves. With the exponential rise in online platforms for goods, services, and social networking, users often encounter overwhelming amounts of content, making effective filtering essential (Panzer and Gronau 2024; Liang et al. 2023). Recommendation systems have thus become a staple on such platforms, and the proposed model offers a robust, node-centered approach to meet these demands. By adapting to preferences and the nature of connections within social graphs-representing friendships, trust links, or shared tastes-CyS delivers dynamic, adapt­ able recommendations across a range of applications (Heshmati et al. 2025). 1.1  Formalization of the social graph structure A social graph is defined as G = (U, I, E), where: 1. U = {u1, u2, . . . , un} represents the set of nodes in the graph, where each

This content is AI-processed based on open access ArXiv data.