Human Immunodeficiency Virus-1 protease (HIV-1 PR) is among the most extensively studied drug targets in the Protein Data Bank (PDB), with more than 600 structural models predominantly derived by X-ray crystallography. This study presents a comprehensive analysis of the binding-site conformational space across the available structural record: 690 crystal structures deposited in the PDB with ≥90% sequence identity and resolution better than or equal to 2.50 Å, of which 684 were successfully featurized by pipeline. The structural dataset covers wild-type enzyme, crystallographic stabilization mutants, drug-resistant variants, and 452 distinct inhibitor binders. Each binding site was featurized as a volume-filling point cloud with six descriptors (electrostatic potential, lipophilicity, and pharmacophoric features) and represented as a geodesic distance matrix with further embedding in spectral distance space. Affinity propagation clustered all pockets into 16 discrete conformational states, with four dominant states accounting for 85% of all structures.

Adaptive classification of air source heat pump operation modes in different climate zones of plateau based on affinity propagation algorithm

To provide services during large-scale natural disasters, it is crucial for network operators to have adaptive and intelligent solutions. With this in mind, new solutions need to be developed, as conventional ground base stations (GBSs) may not be suitable or fast enough to provide services in such emergency situations. Hence, a research gap for Emergency Communications Networks (ECNs) is the deployment of unmanned aerial vehicles (UAVs) in emergency areas. To address this research gap, this paper focuses on the efficient and optimal three-dimensional deployment of UAVs in scenarios characterized by high user density and heterogeneous distributions. The purpose of this study is to address real-world challenges, including the spatial distribution of users and the simultaneous presence of multiple GBSs. This study attempted to develop a novel data clustering approach for wireless networks, based on the Affinity Propagation algorithm, referred to as Deep-Embedded Affinity Propagation Clustering (DEAPC). By integrating deep learning to map data into a latent feature space, this method enhances the clustering algorithm’s capability to handle scattered and noisy data. Moreover, a novel mechanism is designed to accommodate the presence of multiple GBS and to assign users to them. Simulation outcomes demonstrate that the new approach outperforms current leading clustering algorithms, reducing the required number of UAVs whereas also increasing system sumrate and decreasing computational time. This research presents a new method for creating intelligent, resilient, and adaptive UAV-based wireless networks in disaster scenarios.

Machine learning techniques provide powerful tools for analysis of encrypted network traffic. We present a novel set of features and a distance measure suitable for a wide variety of distance-based machine learning techniques useful in classification, clustering, and novelty detection in encrypted traffic flow data. The proposed features are given by distances between probability distributions of such characteristics as packet sizes or inter-arrival times. The proposed distance measure incorporates those features in a fractional l p -metric with p close to 0.1. The effectiveness of the distance measure is evaluated using an extensive labeled dataset containing 154 web services. The dataset was captured on the ISP backbone, and we present it as supplementary material. Application of k -nearest neighbors (kNN) algorithm in combination with the proposed distance measure and feature selection gives the classification accuracy of 91.0%. Comparison with a deep learning model shows that the kNN is competitive with state-of-the-art models. On a different publicly available dataset with a low number of classes, our approach reaches accuracy 99.6%, outperforming models presented in the literature. The benefits of novel features are further demonstrated using the LightGBM model, i.e. without relying on distance-based techniques. Besides direct classification, we have used the local outlier factor and rank-based detection algorithms to detect novel traffic flows. We show that their performance improves when using the proposed distance measure. Finally, to speed up traffic classification we compared the performance of kNN and Approximate Nearest Neighbors (ANN) algorithm, and applied the Affinity propagation clustering algorithm to select a suitable subset of kNN/ANN training points.

Pixel-level annotation of lung cavities (LCs) in computed tomography (CT) images is challenging due to their morphological diversity and complexity. Weakly supervised semantic segmentation (WSSS) methods, which utilize sparse annotations (e.g., image-level labels), offer a promising solution. However, existing WSSS approaches often generate coarse pseudo-labels and lack sufficient spatial supervision, resulting in under- or over-segmentation of irregular lesions. To address these limitations, we introduce several key innovations. First, we propose a novel Graph-based Affinity Network (GA-Net) that, unlike conventional methods relying on low-level pixel features, models long-range contextual relationships and structural dependencies using a superpixel graph and learned edge inference kernel, enabling structure-aware pseudo-label refinement for complex lesion morphology. Second, we introduce region-wise affinity propagation, which refines segmentation by propagating activations within semantically coherent 3D regions, offering more precise control over under-/over-segmentation compared to global affinity methods. Additionally, we incorporate Exponential Moving Average (EMA) ensembling for training stability and a scribble-based segmentation module that utilizes pseudo-label contours to provide direct boundary supervision. Extensive experiments on three benchmark datasets demonstrate that our method outperforms existing state-of-the-art medical WSSS techniques, achieving precise and reliable segmentation of complex LCs in CT scans.

Background/Objectives: Accurate segmentation of gliomas in 3D volumetric MRI is critical for diagnosis, treatment planning, and surgical navigation. However, the scarcity of expert annotations limits the applicability of supervised learning approaches, motivating the development of unsupervised methods. This study presents US-ATHC (Unsupervised Segmentation using Adaptive Thresholding and Hierarchical Clustering), a fully unsupervised two-step pipeline for both global tumor detection and multi-class subregion segmentation. Methods: In the first step, a global tumor mask is extracted by combining adaptive thresholding (Sauvola) with morphological processing on individual MRI slices. The resulting candidates are fused across axial, coronal, and sagittal views using a strict 3D consistency criterion. In the second step, the global mask is refined into a three-class segmentation (active tumor, edema, and necrosis) using optimized affinity propagation clustering. Results: The method was evaluated on the BraTS 2021 dataset, demonstrating accurate tumor and subregion segmentation that outperformed both classical clustering techniques and state-of-the-art deep learning models. External validation on the Gliobiopsy dataset from the University Hospital of Poitiers confirmed robustness and practical applicability in real-world clinical settings. Conclusions: US-ATHC establishes an unsupervised paradigm for glioma segmentation that balances accuracy with computational efficiency. Its annotation-independent nature makes it suitable for scenarios with scarce labeled data, supporting integration into clinical workflows and large-scale neuroimaging studies.

Pseudosulfitobacter pseudonitzschiae is a species within the genus Pseudosulfitobacter, which belongs to the Roseobacteraceae. This family is closely associated with algae and is essential to marine ecosystems, particularly through interactions with phytoplankton. Notably, this bacterium can produce bioactive compounds that influence microbial dynamics and algal growth in marine environments. In addition to essential nutritional factors, the marine green macroalgal genus Ulva (Chlorophyta) relies on a combination of regulatory morphogenetic compounds produced by its associated epiphytic bacteria to achieve proper morphogenesis. Since P. pseudonitzschiae is rarely described and phylogenetic clustering within this genus is challenging, we conducted phylogenetic and genomic analyses to better resolve its taxonomic position and to explore its functional potential in the Antarctic environment. P. pseudonitzschiae BPC-C4-2 was isolated and integrated into a tripartite model system alongside Maribacter sp. BPC-D8 (CP128187.1) and Ulva sp. UPC-109 (PP091299.1). This biosystem was designed to study the mechanisms of cold-water adaptation involved in the morphogenetic development of the genus Ulva. The hybrid genome assembly of P. pseudonitzschiae BPC-C4-2 consisted of seven contigs totaling 5,450,390 bp, with a GC content of 59.0%. Genome annotation identified 5,380 coding sequences (CDSs), 6 rRNA genes, 85 tRNA genes, and 1 tmRNA. The relatively large number of coding sequences and RNA genes observed may reflect an expanded genetic toolkit that enables metabolic flexibility and stress tolerance, potentially supporting adaptation to the extreme conditions of Antarctic and cold-water environments. Given the taxonomic complexity within the bacterial family, both 16S rRNA gene sequencing and average nucleotide identity (ANI) analyses were conducted. Using affinity propagation clustering, these analyses enabled a more robust phylogenetic placement of P. pseudonitzschiae within this challenging group, providing deeper ecological and evolutionary insights. In this study, we searched for gene clusters associated with metabolic adaptations. Functionally, the genome harbors a modularly organized sox gene cluster involved in the sarcosine oxidation pathway and dimethylsulfoniopropionate (DMSP) degradation. Additional pathways involved in osmolyte metabolism and methylation were also identified and found to be phylogenetically distinct from closely related species. Our findings provide the basis for a cold-water bioassay system to study the morphogenesis of Ulva collected from polar regions, in association with Maribacter sp. BPC-D8 and P. pseudonitzschiae BPC-C4-2. Genomic analyses of P. pseudonitzschiae BPC-C4-2 provide insights into its revised phylogeny based on affinity propagation clustering, along with a detailed analysis of genes involved in key metabolic pathways.

In the field of logistics distribution, the two-echelon vehicle routing problem has long been a critical focus. Against the backdrop of global warming, enterprises conducting logistics operations must now prioritize not only delivery costs but also the environmental impact of carbon emissions. To address these challenges, this study integrates occasional drivers into the two-echelon vehicle routing framework, centering on carbon emission reduction. First, Affinity Propagation (AP) clustering is applied to assign customer points to transfer centers. Subsequently, an optimization model is formulated to minimize both vehicle routing costs and carbon emission costs through a collaborative delivery system involving specialized and crowdsourced vehicles. An enhanced Sparrow–Whale Optimization Algorithm (S-WOA) is proposed to solve the model. The algorithm is tested against traditional heuristic methods on three datasets of different scales. Experimental results demonstrate that the two-echelon logistics and distribution model combining specialized vehicles and occasional drivers achieves a significant reduction in total delivery costs compared to models relying solely on specialized vehicles. Further analysis reveals that, with a fixed crowdsourced compensation coefficient, increasing the crowdsourced detour coefficient leads to a decline in total delivery costs. Conversely, when the detour coefficient remains constant, raising the compensation coefficient results in an upward trend in total costs. These insights provide actionable strategies for optimizing cost-efficiency and sustainability in logistics operations.

Impurity-based borderline SMOTE with affinity propagation for imbalanced data classification

Scene synthesis is crucial to computer graphics. However, the current interactive scene synthesis methods usually cost the user too much time and interactions to edit objects. This paper presents a new interactive scene synthesis method that alleviates the designer from interacting with the 3D scene and its objects. Designers only need to select an object group in an independent panel through coarse clustering and fine clustering. Then, the furniture objects will be automatically added to the scene. This paper proposes a two-level clustering that applies the Affinity Propagation Algorithm (APA) to groups of furniture objects such that the object groups can even be clustered without linear representations, latent encoding, etc. To fully apply the APA, we also propose quantitatively measuring how different the two layouts are, i.e., how quantitatively the arrangements of two object groups differ. Experiments first show that our method is more user-friendly and interactively efficient than other interactive synthesis methods. By comparing our method with recent automatic scene synthesis methods, we demonstrate that our methods still have competitive plausibility. We also verify that our method does not harm the diversity and generalization of 3D scenes.

Purpose This study explores first-year Electrical Engineering and Computer Science students’ use and perception of artificial intelligence (AI) tools in a programming course, and their preferences for future development. Design/methodology/approach We conducted an anonymous exploratory survey, consisting of items with predefined response options and open-ended items. Responses to the former and open-ended items were analyzed using descriptive statistics and inductive thematic analysis, respectively. Additionally, we clustered students using the Affinity Propagation algorithm based on their expressed preferences for possible improvements of AI tools Findings The findings show that AI tools are not universally effective, with seven student clusters identified based on differing needs and expectations. Students expressed a need for AI tools that offer more detailed error explanations and guidance rather than just delivering correct solutions. The most common concern among students is the provision of correct solutions without adequate explanations of the underlying mistakes, leading to a lack of deeper understanding Originality/value This study takes an exploratory approach by examining students’ perceptions and preferences for the design and capabilities of AI tools in helping them learn programming. Clustering students by preferences reveals distinct approaches that may be needed for different groups of learners. Given the limited research on such desires or on applying clustering to them, our analysis offers valuable insights into distinct viewpoints that can guide the design of future personalized educational AI tools

In indoor positioning systems, it is common to use existing AP deployments within buildings to build a fingerprint database, providing positioning information during the online phase. However, AP layouts inside buildings often contain a large number of redundant APs, which leads to the improvement in positioning accuracy leveling off as the number of redundant APs increases, while also increasing the computational load of indoor positioning services. To address this problem, the thesis proposes a method for calculating the AP location discrimination capability and combines the location discrimination capability with coverage to eliminate redundant APs. Experiments conducted in real indoor scenarios, as well as on the Crowdsourced dataset and the SODIndoorLoc dataset, validate the results. The results show that the redundant AP removing strategy ensures that the average positioning accuracy fluctuates by no more than 5% compared to the unfiltered case, while significantly reducing the number of APs in the fingerprint database—by 64.43%, 72.78%, and 59.62%, respectively. In the position estimation phase, this paper uses affinity propagation clustering for coarse positioning and combines Bayesian methods for fine positioning. Compared with GMM, K-Means, and the pointwise algorithm, the average positioning error of the proposed method is reduced by 11% to 39%.

Using the BIRCH algorithm and affinity propagation, an advanced descriptor for video processing

Bandwidth-intensive applications and User Equipment (UE) mobility demand an adaptive Radio Access Network (RAN) that can detect user hotspots to maintain service quality. Virtual Small Cells (VSCs) address this need via beamforming, yet Cell-Head-based VSCs (CHVSC/UE-VBS) introduce additional hardware requirements and are vulnerable to uncertainty in qualified-UE (qUE) willingness to act as a Cell Head (CH). To bridge this practicality gap, we propose a Mixed VSC (MVSC) framework that switches between direct VSC service (DVSC) and CHVSC operation according to qUE willingness. This mode switching alters channel capacity, requiring RB-feasible member association adjustments—an aspect largely absent from clustering-only VSC studies. We therefore develop Load-Aware Affinity Propagation Clustering (LAPC), which integrates predicted link capacity and RB demand into a greedy refinement process that iteratively adjusts cluster membership to mitigate persistent underload/overload across DVSC and CHVSC modes. Monte Carlo simulations compare LAPC against K-Means, MAPC, and GAPC-SNR. Results show that LAPC improves RB utilization while serving more UEs and increasing aggregate throughput, achieving mean RB-utility gains of 8.68 and 10.66 percentage points over K-Means and GAPC-SNR, respectively, in the evaluated MVSC scenario.

ABSTRACT With the rapid increase in penetration of distributed generation predominantly based on renewable energy sources within distribution networks, the inherent volatility and intermittency of distributed generation exacerbate voltage fluctuations and uneven voltage profiles. Concurrently, insufficient coordination mechanisms for diverse controllable devices lead to frequent limit violations of automatic voltage control actions at substations, causing significant voltage control challenges. To address these issues, this study proposes a reactive power voltage control method for distribution networks based on a ‘network partitioning and distributed scheduling’ process. First, the affinity propagation clustering algorithm is employed to partition the network based on the reactive power‐voltage sensitivity matrix. Subsequently, a Markov game model is formulated for voltage control within each partition and the twin delayed deep deterministic policy gradient algorithm is utilised for a distributed solution, outputting target reactive power setpoints for both distributed generators and static Var compensators. Compared to conventional real‐time control methods, the proposed approach significantly reduces the dependence on centralised communication infrastructures and achieves autonomous regional control of reactive power and voltage. Consequently, it effectively minimises the frequency of automatic voltage control corrective actions and enhances the robustness of voltage control. Finally, the effectiveness of the proposed methodology is validated through simulation case studies utilising the IEEE 33‐bus test system and a real‐world network model based on the Beijing distribution grid.

Graph variational autoencoder with affinity propagation for community-aware anomaly detection in attributed networks

Pixel-level annotation of lung cavities (LCs) in computed tomography (CT) images is challenging due to their morphological diversity and complexity. Weakly supervised semantic segmentation (WSSS) methods, which utilize sparse annotations (e.g., image-level labels), offer a promising solution. However, existing WSSS approaches often generate coarse pseudo-labels and lack sufficient spatial supervision, resulting in under- or over-segmentation of irregular lesions. To address these limitations, we introduce several key innovations. First, we propose a novel Graph-based Affinity Network (GA-Net) that, unlike conventional methods relying on low-level pixel features, models long-range contextual relationships and structural dependencies using a superpixel graph and learned edge inference kernel, enabling structure-aware pseudo-label refinement for complex lesion morphology. Second, we introduce region-wise affinity propagation, which refines segmentation by propagating activations within semantically coherent 3D regions, offering more precise control over under-/over-segmentation compared to global affinity methods. Additionally, we incorporate Exponential Moving Average (EMA) ensembling for training stability and a scribble-based segmentation module that utilizes pseudo-label contours to provide direct boundary supervision. Extensive experiments on three benchmark datasets demonstrate that our method outperforms existing state-of-the-art medical WSSS techniques, achieving precise and reliable segmentation of complex LCs in CT scans.

In the active distribution networks (ADNs) with high proportion photovoltaic (PV) access, current coordinated control methods based on real‐time operating data or forecast data cannot meet the requirements of cost minimization and accuracy maximization at the same time. First, the traditional PV probability density function is optimized by a random simulation method. The expected matrix of electric distance between nodes is obtained by discretizing the characteristics of PV output probability distribution. The ADN system is partitioned using the Affinity Propagation (AP) clustering algorithm. On this basis, an upper‐layer scheduling method aiming at minimizing the operating cost of the ADN system is proposed, and day‐ahead scheduling plans are developed for devices such as on‐load tap changer (OLTC) and capacitor banks (CBs) in each cluster. At the lower layer, real‐time operational data are utilized to develop cluster voltage control strategies with the goal of maximizing the tracking of day‐ahead scheduling plans, ensuring the economic and safe operation of the ADN while compensating for the shortcomings caused by forecast errors. The proposed method is tested on IEEE‐123‐bus ADN systems. The simulation results fully verify that the control strategy described in this paper has faster solving speed, better optimization effect, and more feasibility. © 2025 Institute of Electrical Engineers of Japan and Wiley Periodicals LLC.

Classifying organometallic reactions into distinct reaction types is fundamentally important for understanding mechanisms and predicting reactions, for synthesis optimisation. Fundamental to classification of organometal reactions is reaction representation, but current methods often fail to capture organometal chemical transformation adequately. The study has adopted a hybrid fingerprinting approach, whereby new fingerprints were concatenated with permutation important Morgan fingerprints to create 49 to 63 bits fingerprints. The hybrid fingerprints were used to build KMeans clustering, Guassian mixtures, Affinity propagation and Heirarchical clustering models for classification of organometal reactions. The models successfully classified reactions into 6-8 major organometal reaction types. Of note, the fingerprints consistently outperformed Morgan fingerprints across all clustering models with good Davies–Bouldin Index (DBI) ranging from 0.3 to 0.6 and Silhouette score from 0.3 to 0.8. Furthermore, the clustering models were visualized using Principal Component Analysis. Affinity Propagation and KMeans demonstrated superior performance over Hierarchical and GMM algorithms in distinguishing major reaction categories. In contrast, the Hierarchical model excelled at identifying sub-level classifications compared to the other methods. Consequently, Affinity Propagation and KMeans are recommended for broad reaction type classification, while the Hierarchical approach is better suited for resolving detailed subclass distinctions. The observed variability in model performance further highlighted the importance of feature selection and representation in clustering models. Future studies should look into improving the fingerprints to recognize subcategories.

As computers develops, virtual simulation technology becomes an important means of integrated circuit design. Therefore, based on the demand for virtual simulation of integrated circuits, a simulation method combining affinity propagation and differential evolution algorithm was proposed. By applying the affinity propagation to circuit fault diagnosis and combining it with differential evolution algorithm, circuit parameters optimization was carried out. These experiments confirm that the fusion of affinity propagation and differential evolution algorithm has a precision of 94.26%, recall of 93.41%, mean F1 of 88.59%, convergence speed of 56.77 seconds, and stability of 93.17%. The affinity propagation performs well in clustering. Especially without pre-defining the classes, it can identify the position and number of class centers automatically. The simulation of integrating affinity propagation and differential evolution algorithm has broad application prospects in virtual simulation of integrated circuits. It can improve simulation effectiveness and performance, providing effective support for circuit design and testing.