Unsupervised domain adaptation (UDA) remains challenged by an unstable target structure, pseudo-label noise, and heterogeneous transfer difficulty across domains. To address these issues, we propose Progressive Multi-View Graph Projection (PMGP), a two-stage framework that first learns transferable representations via source supervision, domain-adversarial training, and teacher–student consistency and then performs latent-space refinement through multi-view graph construction and projection learning. Specifically, three perturbation-induced views are considered for each sample: the original view, an input-space patch-masked view, and a representation-space feature-dimension masked view. After joint preprocessing with PCA and L2 normalization, PMGP constructs per-view affinity graphs by combining geometric neighborhood relations with pseudo-supervised semantic relations, and applies locality-preserving projection to learn a structure-aware shared subspace. In this subspace, target pseudo-labels are iteratively refined using source prototypes, target class centers, and progressive confidence filtering. Experiments on Office-Home, ImageCLEF-DA, and VisDA-2017 show that PMGP achieves competitive performance and stable behavior across different benchmark settings and backbone architectures. These results indicate that multi-view graph refinement provides an effective and interpretable way to improve target structure estimation and reduce pseudo-label error accumulation in UDA.

Band selection plays a critical role in reducing data redundancy of hyperspectral images, where subspace clustering-based methods have shown remarkable potential, owing to the effective extraction of low-dimensional representations. Despite conducting band grouping competently, they not only overlook preserving original affinity structures but also require a manual graph normalization step, which could lead to suboptimal performance. To solve these issues, we propose a novel band selection method termed Implicit Doubly Stochastic Graph based subspace Clustering (IDSGC), facilitating end-to-end band affinity graph construction. Firstly, a latent bipartite graph modeling (LBGM) module is designed to decompose the inherent band similarities into bipartite graph, while mitigating the interference of noise links. Secondly, a novel implicit learning mechanism (ILM) of doubly stochastic graph is presented to refine graph structure and low-dimensional representations simultaneously, which seamlessly integrates the LBGM with subspace learning. Finally, a probability-constrained affinity graph that is implicitly normalized during optimization is generated, which can be directly used for band clustering without additional post-processing. Extensive experiments demonstrate that IDSGC outperforms several state-of-the-art band selection methods.

Lifting multi-view 2D masks generated by the Segment Anything Model (SAM) into 3D space offers a promising direction for zero-shot 3D scene segmentation, but view-dependent occlusions and limited fields of view often cause incomplete observations and cross-view inconsistencies, resulting in fragmented semantics and geometric misalignment. To address this, we propose SAM-Zero3D, which extends SAM to the 3D domain through a structured fusion pipeline with two complementary branches. The global anchor point-guided branch projects 3D anchors into multi-view masks to construct a cross-view affinity graph, identifies consistent mask groups via connected component analysis, and assigns 3D masks via majority voting and nearest-neighbor propagation. The local geometry-driven branch partitions the point cloud into fine-grained regions, estimates region-level semantic similarity from aggregated mask distributions, and progressively merges similar regions through a multi-stage merging strategy. An iterative global–local interaction further refines both branches by aligning global semantic priors with local geometric cues. Extensive experiments on ShapeNetPart, ScanNetV2, and ScanNet200 show that SAM-Zero3D significantly outperforms existing zero-shot baselines, achieving accurate and structure-aware segmentation without any 3D training or supervision.

Existing multi-view clustering methods often separate feature learning from clustering, leading to suboptimal preservation of global structure and increased reliance on heuristic post-processing. Recently risen tensor-based approaches better model cross-view correlations but still fail to fully capture intrinsic sample relationships. To tackle these shortcomings, we propose a unified one-step multi-view clustering framework (MCOC) that combines contrastive learning, tensor-based self-representation, and nonnegative symmetric matrix factorization. View-specific self-representation matrices are obtained via contrastive learning and arranged along the third mode to form a third-order tensor to capture global consensus across multiple views, then view-specific self-representation matrices are symmetrized to construct affinity graphs. Joint factorization of degree-normalized affinity matrices yields a shared nonnegative indicator matrix, enabling direct cluster assignment. Experiments on real-world datasets validate the superior clustering performance of the proposed method, effectively fusing local view-specific and global structural information.

This paper investigates the problem of dynamical sampling for graph signals influenced by a constant source term. We consider signals evolving over time according to a linear dynamical system on a graph, where both the initial state and the source term are bandlimited. We introduce two random space-time sampling regimes and analyze the conditions under which stable recovery is achievable. While our framework extends recent work on homogeneous dynamics, it addresses a fundamentally different setting where the evolution includes a constant source term. This results in a non-orthogonal-diagonalizable system matrix, rendering classical spectral techniques inapplicable and introducing new challenges in sampling design, stability analysis, and joint recovery of both the initial state and the forcing term. A key component of our analysis is the spectral graph weighted coherence, which characterizes the interplay between the sampling distribution and the graph structure. We establish sampling complexity bounds ensuring stable recovery via the Restricted Isometry Property (RIP), and develop a robust recovery algorithm with provable error guarantees. The effectiveness of our method is validated through extensive experiments on both synthetic and real-world datasets.

Multi-graph clustering via multi-modal topological manifold learning.

One-step multi-view kernel clustering based on topological manifold

PEARL: A dual-layer graph learning for multimodal recommendation

This study integrates the Kano model and affinity graph to conduct in-depth exploration of the demand characteristics of regional embroidery products. Through in-depth interviews and questionnaire surveys, 15 key demand elements were extracted from four dimensions: design, production, cultural connotation, and user experience. Using the Kano model classification and Better Worse coefficient calculation, it was found that five requirements, including "cultural heritage" and "practicality," belong to essential qualities (M), four requirements, including "fashionability" and "traditional craftsmanship," belong to one-dimensional qualities (O), and five requirements, including "personalized customization" and "artistic quality," belong to charm qualities (A). Among them, "collectability" has the highest sensitivity (ω=1.145), and "bright colors" and other four items belong to indifference quality (I). Based on this, suggestions are proposed to focus on the demand for charm, innovate design, and strengthen cultural narrative through the Internet, providing scientific basis for product differentiation development and precise market positioning.

Towards Dynamic Self-Training for Scalable Semi-Supervised Learning on Graphs

Identifying accurate cell markers in single-cell RNA-seq data is crucial for understanding cellular diversity and function. Localized Marker Detector (LMD) is a novel tool to identify “localized genes”—genes exclusively expressed in groups of highly similar cells—thereby characterizing cellular diversity in a multi-resolution and fine-grained manner. LMD constructs a cell-cell affinity graph, diffuses the gene expression value across the cell graph, and assigns a score to each gene based on its diffusion dynamics. LMD’s candidate markers can be grouped into functional gene modules, which accurately reflect cell types, subtypes, and other sources of variation such as cell cycle status. We apply LMD to mouse bone marrow and hair follicle dermal condensate datasets, where it facilitates cross-sample comparisons by identifying shared and sample-specific gene signatures and novel cell populations, without requiring batch effect correction or integration. We also assess the performance of LMD across ten single-cell RNA sequencing datasets, compare it to eight existing methods with similar objectives, and find that LMD outperforms the other methods evaluated.

IntroductionPremature Ventricular Contractions (PVCs) can be warning signs for serious cardiac conditions, and early detection is essential for preventing complications. The use of deep learning models in electrocardiogram (ECG) analysis has aided more accurate and efficient PVC identification. These models automatically extract and analyze complex signal features, providing valuable clinical decision-making support. Here, we conducted a study focused on the practical applications of is technology.MethodsWe first used the MIT-BIH arrhythmia database and a sparse low-rank algorithm to denoise ECG signals. We then transformed the one-dimensional time-series signals into two-dimensional images using Markov Transition Fields (MTFs), considering state transition probabilities and spatial location information to comprehensively capture signal features. Finally, we used the BiFormer classification model, which employs a Bi-level Routing Attention (BRA) mechanism to construct region-level affinity graphs, to retain only the regions highly relevant to our query. This approach filtered out redundant information, and optimized both computational efficiency and memory usage.ResultsOur algorithm achieved a detection accuracy of 99.45%, outperforming other commonly-used PVC detection algorithms.DiscussionBy integrating MTF and BiFormer, we effectively detected PVCs, facilitating an increased convergence between medicine and deep learning technology. We hope our model can help contribute to more accurate computational support for PVC diagnosis and treatment.

Recommender systems are widely used in various real-world applications, but they often encounter the persistent challenge of the user cold-start problem. Cross-domain recommendation (CDR), which leverages user interactions from one domain to improve prediction performance in another, has emerged as a promising solution. However, users with similar preferences in the source domain may exhibit different interests in the target domain. Therefore, directly transferring embeddings may introduce irrelevant source-domain collaborative information. In this paper, we propose a novel graph-based disentangled contrastive learning framework to capture fine-grained user intent and filter out irrelevant collaborative information, thereby avoiding negative transfer. Specifically, for each domain, we use a multi-channel graph encoder to capture diverse user intents. We then construct the affinity graph in the embedding space and perform multi-step random walks to capture high-order user similarity relationships. Treating one domain as the target, we propose a disentangled intent-wise contrastive learning approach, guided by user similarity, to refine the bridging of user intents across domains.Extensive experiments on four benchmark CDR datasets demonstrate that DisCo consistently outperforms existing state-of-the-art baselines, thereby validating the effectiveness of both DisCo and its components.

Kernelized multi-view graph clustering via graph structure preserving and consensus affinity graph learning

Computational methods for predicting drug-target binding affinity (DTA) are critical for large-scale screening of prospective therapeutic compounds during drug discovery. Deep neural networks (DNNs) have recently shown significant promise for DTA prediction. By leveraging available data for training, DNNs can expand the use of DTA prediction to situations where only sequence information is available for potential drug molecules and their targets, and there is no prior knowledge regarding the molecular geometric conformations. We propose DHAG-DTA, a general dynamic hierarchical affinity graph DNN approach, for DTA prediction using molecular sequence information and already known drug-target interactions. DHAG-DTA introduces a two-level hierarchical graph structure: at the upper level, interactions between drug and target molecules are represented via an affinity graph and at the lower level, embedded molecular graphs represent interactions within the individual molecules. This allows for integration of information from both inter and intra molecular interactions for DTA prediction, which has also been addressed in other recent independent work. The fundamental innovations introduced by DHAG-DTA include: (a) a single overall hierarchical graph that allows better assimilation of information during the learning process compared with loosely-coupled individual graphs, (b) dynamic determination of the affinity graph structure via the introduction of unlabeled edges and a maximum entropy criterion for active edge selection, (c) skip connections in the DNN for fusing intra and inter molecular information, and (d) fusion of both model-based and similarity-based feature embeddings to get robust embeddings of unseen molecules. Experimental results on two common benchmark datasets demonstrate that DHAG-DTA outperforms other existing models on multiple evaluation metrics, achieving state-of-the-art performance.

In graph based multiview clustering methods, the ultimate partition result is usually achieved by spectral embedding of the consistent graph using some traditional clustering methods, such as -means. However, optimal performance will be reduced by this multistep procedure since it cannot unify graph learning with partition generation closely. In this article, we propose a one-step multiview clustering method through adaptive graph learning and spectral rotation (AGLSR). For every view, AGLSR adaptively learns affinity graphs to capture similar relationships of samples. Then, a spectral embedding is designed to take advantage of the potential feature space shared by different views. In addition, AGLSR utilizes a spectral rotation strategy to obtain the discrete clustering labels from the learned spectral embeddings directly. An effective updating algorithm with proven convergence is derived to optimize the optimization problem. Sufficient experiments on benchmark datasets have clearly demonstrated the effectiveness of the proposed method in six metrics. The code of AGLSR is uploaded at https://github.com/tangchuan2000/AGLSR.

Multi-view graph clustering (MVGC) utilizes affinity graphs to efficiently obtain information between views. Although various excellent MVGC methods have been proposed, they still have many limitations. To surmount these limitations, this work develops a novel tensor-based unified and discrete multi-view projection clustering (TUDMPC) approach. Specifically, TUDMPC uses projection and the L2,1-norm for feature selection to reduce the effects of redundancy and noise. Meanwhile, the differences among similar graphs are minimized through the tensor kernel norm to better leverage information across views and capture high-order correlations. In addition, the rank constraint is applied to keep the affinity graphs with a discrete cluster structure, and the clustering results are obtained directly in a unified joint framework. Finally, an efficient optimization algorithm is proposed to obtain the clustering results. Experiments are conducted to compare the clustering results of TUDMPC with seven baseline methods. The results show that TUDMPC outperforms the existing methods.

ABSTRACT Denoising is a key preprocessing to enhance contaminated images. There have been lots of literature tackling such an issue. Most of them focus on pixel‐level processing and ignore intrinsic sparsity and low‐rank properties of noise and objects, respectively, in images. To argue this issue, a novel image enhancement model is proposed to eliminate or suppress noise, which integrates robust principal component analysis with kernel dictionary learning (KDL). Specifically, the Kernel trick is introduced to nonlinearly map image data and low‐rank dictionary to be learned to high‐dimensional spaces so as to separate sparsity noise from contaminated images. The proposed model is abbreviate as RSKDL. In RSKDL, the intrinsic structural characteristic of images is unclosed by adaptively learning the affinity graph of image data so as to ensure the enhanced images inheriting the manifold structure of original images. Meantime, the non‐convex sparsity regularization on the residual between original images and enhanced ones is imposed to exclude noise from original data. Extensive experiments on several image datasets show that the proposed model outperforms existing methods for image denoising.

The accurate analysis of the cost data of power grid project has important construction guiding significance.There are many analyses on the influence of power engineering cost. Therefore, a method of cost analysis of power engineering based on affinity graph is proposed in this paper. By weighting the factors that affect the cost of electric power engineering and classifying the weights of each factor, the method can effectively extract the main index factors that affect the cost of electric power engineering. In addition, the support vector machine (SVM) model is used to analyze and evaluate the cost data. The results show that the affinity graph method used in this study can effectively evaluate the cost of electric power engineering, and the algorithm can effectively simplify the human operation, and further improve the evaluation efficiency.

Multiview diffusion-based affinity graph learning with good neighbourhoods for salient object detection