Image clustering aims to partition unlabeled image datasets into distinct groups. A core aspect of this task is constructing and leveraging prior knowledge to guide the clustering process. Recent approaches introduce semantic descriptions as prior information, most of which typically relying on matching-based techniques with predefined vocabularies. However, the limited matching space restricts their adaptability to downstream clustering tasks. Moreover, these methods primarily focus on reducing bias to improve performance, frequently overlooking the importance of variance reduction. To address these limitations, we propose GSEC (Image Clustering based on Generative Semantic Guidance and Bi-Layer Ensemble), a framework designed to reduce bias through generative semantic guidance and mitigate variance via ensemble learning. Our method employs Multimodal Large Language Models to generate semantic descriptions and derive image embeddings via weighted averaging. Additionally, a bi-layer ensemble strategy integrates cross-modal information through BatchEnsemble in the inner layer and aligns outputs via an alignment mechanism in the outer layer. Comparative experiments demonstrate that GSEC outperforms 20 state-of-the-art methods across six benchmark datasets, while further analysis confirms its effectiveness in simultaneously reducing both bias and variance.

With the growing demand for decentralized collaborative analysis of privacy-sensitive data, federated multi-view clustering (FMVC) has attracted widespread attention due to its ability to balance privacy protection and collaborative modeling. However, current methods still face the following challenges: (1) Clients need to frequently upload high-dimensional data such as model parameters or graph structures, resulting in high communication costs; (2) The structured data uploaded often contains semantic features and has a high risk of being inverted; (3) The server usually merges the data from all clients with the fixed fusion rule, which may result in a suboptimized clustering result when there exist low-quality clients. To address the issues, we propose a new trusted federated multi-view clustering framework (EvoFMVC) that introduces three key innovations: First, lightweight trusted evidence serves as a compact communication medium, significantly reducing overhead compared to conventional model parameters or graph structures. Second, trusted evidences express clustering results in the form of probability distribution, which avoids the risk of structured information being easily inverted. Lastly, we formalize the server-side aggregation process as a neural architecture search (NAS) task where the server flexibly uses different fusion operators to filter and fuse necessary views through evolutionary algorithms, which significantly improves the fusion effect and model performance. Experimental results on multiple datasets show that our method is superior to existing FMVC methods in terms of clustering accuracy and communication efficiency.

Multi-view clustering seeks to leverage complementary information from heterogeneous data sources, yet existing methods often overlook two critical aspects: per-cluster view relevance and fine-grained intra-view structures. We propose a novel Dual Mixture-of-Experts (Dual MoE) framework for multi-view K-means that simultaneously addresses these challenges. First, our cluster-view adaptive weighting dynamically calibrates each view's importance at the cluster level, enabling the model to emphasize informative views while suppressing noise. Second, our region-wise sparse partitioning decomposes each view into localized subregions, capturing nuanced intra-view patterns via specialized centroids. We derive high-probability excess clustering risk bounds, revealing that balanced view-cluster weights minimize generalization error and motivate our entropy-regularized square-weight mapping. A linear-time alternating optimization algorithm ensures scalability to large datasets. Empirical evaluations on ten public benchmarks demonstrate that Dual MoE consistently outperforms state-of-the-art methods, achieving an average rank of 1.1-1.2 across standard accuracy, NMI, and ARI metrics. Ablation studies further confirm the complementary benefits of our dual expert mechanisms. The implementation is available at https://github.com/sroytik/DMoE-MVKM.

In multi-view classification tasks (MVC), each view provides an unique perspective on the data, offering complementary information that can improve classification performance when properly integrated. However, traditional methods typically adopt a uniform processing strategy for all views before fusion, overlooking the fact that different views may require different treatments due to variations in their quality and informativeness. To address this limitation, we propose a novel framework called Uncertainty-Guided View-Strength-Aware Feature Utilization (UVF) for multi-view classification. Our approach introduces a view uncertainty estimation module to quantify the discriminative strength of each view. Based on this estimation, a Differentiated Feature Selector (DFS) adaptively selects features, retaining informative dimensions in weak views while preserving original features in strong views. Furthermore, we employ an uncertainty-guided fusion strategy that assigns dynamic weights to each view's contribution based on its uncertainty score, enhancing the robustness and reliability of the final decision. Experimental results on benchmark datasets demonstrate that our method significantly outperforms conventional approaches, achieving better classification accuracy and interpretability through strength-aware feature processing and fusion.

Similarity-graph-based multi-view clustering is effective in capturing non-Gaussian cluster structures, yet it suffers from an inherent conflict between scalability and geometry-consistent cluster assignment. In particular, cluster labels are typically inferred via secondary clustering rather than being learned directly from the underlying data geometry. Although anchor-based extensions alleviate the computational burden, they introduce the challenge of anchor-number selection. To address these limitations, we propose Multi-view Clustering via Manifold Decomposition (MvMD), which directly infers cluster labels from multi-view data without explicit similarity graph construction or anchor selection. MvMD reformulates multi-view clustering as a unified multi-view regression problem, where cluster labels are optimized as model variables. To improve robustness against missing small clusters, a global balance regularization is incorporated. Meanwhile, local geometric consistency and the low-rank structure of cluster assignments are jointly enforced through a symmetric matrix factorization scheme, which is efficiently realized via a truncated SVD-based low-rank approximation. Extensive experiments on benchmark datasets validate the effectiveness and efficiency of the proposed MvMD. The code is available at https://github.com/Vince-Doit/MvMD.

Multi-view data is widely present in the real world. Multi-view clustering is an unsupervised method for capturing the grouping structure of such data. However, multi-view clustering struggles to meet the requirements of real-world scenarios, such as distributed storage of different views and data protection needs. These requirements align with the setting of vertical federated clustering. However, vertical federated clustering still faces two challenges: (1) Under the constraints of privacy protection mechanisms, how to theoretically analyze the clustering consistency between the data uploaded by clients to the server and the original client data is challenging. (2) The feature space differences among different clients make cross-view information sharing and fusion difficult. To address the first challenge, we provide a theoretical analysis of the upper bound of the loss of k-means for transformation matrix mapping, revealing the relationship between the k-means loss of the transformed data and the original data. We then propose a vertical federated clustering method (V-HDKM). In this method, clients handle the second challenge by transposing the feature matrix. Guided by the projected k-means loss bound, we expand the feature space and perform k-means clustering to obtain feature cluster centers, which are then uploaded to the server. The server aggregates the global centers and feeds back the optimized results, achieving crossview knowledge fusion through iterative interactions. Experimental results show that V-HDKM significantly improves local clustering performance and performances better than other seven vertical federated mthods on 20 multi-view datasets. Furthermore, sensitivity analysis on 8 UCI datasets with respect to the number of clients demonstrates the stability of the method. The code is available at https://github.com/jiangjh/V-HDKM.

Concept factorization provides an interpretable route to clustering by expressing latent concepts as combinations of observed samples. Its development in multiple kernel settings, however, remains limited when one simultaneously requires kernel validity, low storage, and mathematically transparent optimization. This paper proposes a sparse multiple kernel concept factorization framework with adaptive orthogonal factors. The framework first constructs sparse positive semi-definite kernel matrices through localized neighborhood regression and then learns a shared nonnegative consensus representation together with kernel-specific orthogonal factors and adaptive kernel weights. The resulting block coordinate procedure contains two closed-form subproblems and one standard multiplicative update. The positive semi-definiteness of the sparse kernels is proved, the exact solutions of the orthogonality-constrained and weight-update blocks are derived, monotonic descent, lower boundedness, and the limiting behavior of the generated objective sequence are analyzed. The computational and storage complexity of the framework are also analyzed. Experiments on nine benchmark datasets, whose sample sizes range from 103 to 2.8 & times;105 , show that the proposed method remains competitive in clustering quality while retaining attractive runtime and memory behavior. The paper therefore offers a mathematically explicit and empirically verifiable treatment of sparse multiple kernel concept factorization.

Despite significant progress in Visual-Language-Action (VLA), in highly complex and dynamic environments that involve real-time unpredictable interactions (such as 3D open worlds and large-scale PvP games), existing approaches remain inefficient at extracting action-critical signals from redundant sensor streams. To tackle this, we introduce MAIN-VLA, a framework that explicitly Models the Abstraction of Intention and eNvironment to ground decision-making in deep semantic alignment rather than superficial pattern matching. Specifically, our Intention Abstraction (IA) extracts verbose linguistic instructions and their associated reasoning into compact, explicit semantic primitives, while the Environment Semantics Abstraction (ESA) projects overwhelming visual streams into a structured, topological affordance representation. Furthermore, aligning these two abstract modalities induces an emergent attention-concentration effect, enabling a parameter-free token-pruning strategy that filters out perceptual redundancy without degrading performance. Extensive experiments in open-world Minecraft and large-scale PvP environments (Game for Peace and Valorant) demonstrate that MAIN-VLA sets a new state-of-the-art, which achieves superior decision quality, stronger generalization, and cutting-edge inference efficiency.

Clustering ensembles combine multiple base partitions to enhance the robustness, stability, and quality of unsupervised learning. Existing surveys have provided valuable overviews of ensemble generation and consensus functions, but they typically focus on algorithmic taxonomies and pay less attention to how data characteristics shape design choices. At the same time, modern applications increasingly involve large-scale, high-dimensional, multi-view, and dynamic data streams, where naive reuse of classical ensemble techniques may fail. In this survey, we revisit clustering ensembles from a data aware perspective. We organize the literature around a pipeline-based framework with four stages: generation of base clusterings, selection of base clusterings, ensemble representation, and consensus fusion. For each stage, we discuss how key data characteristics, including data type and representation, scale and structure, supervision and external signals, as well as dynamics and resource constraints, affect algorithm design. We further provide representative choices of ensemble components across common data regimes and highlight open challenges for building more data aware clustering ensembles.

Ensemble clustering aims to derive a more robust consensus result from a set of base clustering results. Recently, anchor-based clustering methods have improved computational efficiency by learning the relationships between anchors and samples, thereby avoiding the expensive computation of pairwise sample similarities. However, these methods typically rely on a static anchor selection strategy and require post-processing to obtain the final clustering assignments. This often leads to inflexible anchors and a low-quality anchor similarity matrix, ultimately degrading clustering accuracy. To address this issue, we propose a novel ensemble clustering method named $\mathbf{D}$ynamic ancho $\mathbf{R}$-based One-step hyp$\mathbf{E}$ rgr$\mathbf{A}$ph ense $\mathbf{M}$ ble clustering (DREAM). Specifically, DREAM first transforms the base clustering results into a hypergraph. It then introduces a novel hypergraph enhancement strategy to improve the reliability of the initial hypergraph. Next, DREAM introduces a mapping matrix to filter redundant information in the hypergraph, and reconstructs the hypergraph via matrix factorization to obtain the anchor similarity matrix. Subsequently, DREAM introduces an alignment mechanism that refines the anchor similarity matrix by generating local consensus information. The alignment is achieved by maximizing the element-wise consistency between the anchor similarity matrix and the local consensus information. This yields a high-quality anchor similarity matrix that can be directly projected into the label space, enabling one-step generation of clustering results without any additional post-processing. Extensive experimental results demonstrate the superior performance of the proposed DREAM method.

The Mixture of Experts (MoE) paradigm has been successfully integrated into Low-Rank Adaptation (LoRA) for parameter-efficient fine-tuning (PEFT), delivering performance gains with minimal parameter overhead. However, a key limitation of existing MoE-LoRA methods is their reliance on a discrete router, which prevents the integration of the MoE components into the backbone model. To overcome this, we propose FURINA, a novel Free from Unmergeable Router framework based on the LINear Aggregation of experts. FURINA eliminates the router by introducing a Self-Routing mechanism. This is achieved through three core innovations: (1) decoupled learning of the direction and magnitude for LoRA adapters, (2) a shared learnable magnitude vector for consistent activation scaling, and (3) expert selection loss that encourages divergent expert activation. The proposed mechanism leverages the angular similarity between the input and each adapter's directional component to activate experts, which are then scaled by the shared magnitude vector. This design allows the output norm to naturally reflect the importance of each expert, thereby enabling dynamic, router-free routing. The expert selection loss further sharpens this behavior by encouraging sparsity and aligning it with standard MoE activation patterns. We also introduce a shared expert within the MoE-LoRA block that provides stable, foundational knowledge. To the best of our knowledge, FURINA is the first router-free, MoE-enhanced LoRA method that can be fully merged into the backbone model, introducing zero additional inference-time cost or complexity. Extensive experiments demonstrate that FURINA not only significantly outperforms standard LoRA but also matches or surpasses the performance of existing MoE-LoRA methods, while eliminating the extra inference-time overhead of MoE.

Deep models in dynamic environments often face catastrophic forgetting, where learning new classes degrades recognition of old ones. Class-Incremental Learning (CIL) mitigates this issue, but many methods rely on experience replay, leading to storage overhead and class imbalance. We propose an implicit semantic augmentation framework that operates at the loss-function level. By adaptively estimating class-conditional covariance, our method enhances old-class representations without replay and alleviates imbalance. In addition, we integrate Low-Rank Adaptation (LoRA) for parameter-efficient fine-tuning, where most pre-trained parameters remain frozen and only a few low-rank components are updated. Experiments on multiple vision benchmarks show that our approach achieves competitive performance, demonstrating improved resistance to forgetting. These results indicate that the framework demonstrates strong potential for continual learning and is well-suited for applications such as acoustic event detection and other signal processing tasks. Code is available at: https://github.com/TaoYu-ML/LISA.

Ensemble clustering aims to derive a consensus partition from multiple base clustering results. Anchor-based methods construct compact similarity representations via anchors, substantially improving computational efficiency. However, when outliers contaminate the data, reconstructing the base clustering results often yields biased anchors. These biased anchors degrade the quality of the anchor similarity matrix and lead to a decline in clustering accuracy. To address this issue, we propose a novel method called large-scale robust enhanced ensemble clustering via outlier decoupling (RANGE). Specifically, RANGE first converts the base clustering results into an initial bipartite graph. To enhance the reliability of this bipartite graph, RANGE designs a high-order fuzzy enhancement strategy (HFES) specifically for initial bipartite graphs. Next, a mapping matrix further filters redundant information from the enhanced bipartite graph. RANGE then reconstructs the mapped bipartite graph via matrix factorization. An anchor matrix is introduced to further enhance computational efficiency. To improve robustness, RANGE incorporates a decoupling term that separates the clean clustering structure and the outlier-contaminated structure in the anchor space. With this decoupling mechanism, RANGE is capable of performing robust ensemble clustering. Moreover, by applying outlier detectors to the decoupled outlier structure, RANGE can be extended to the outlier-detection task. Consequently, RANGE forms a cross-task general framework, and both tasks retain linear time complexity. Extensive cross-domain experiments indicate that RANGE delivers superior performance in both clustering validity and outlier detection. The code is available in the supplementary material.

Ensemble clustering integrates a set of base clustering results to obtain more accurate consensus results. Most existing methods transform base clustering results into a co-association (CA) matrix, which describes the frequency with which samples belong to the same cluster. However, these methods still exhibit insufficient learning of sample similarity relationships, overlooking the capture of a more comprehensive similarity relationship from multiple perspectives. To address this issue, we propose a method called tOpological-aware Multiple kErnel Learning for Ensemble clusTering (OMELET). Specifically, OMELET uses multiple kernel functions to remeasure the similarity between samples from multiple perspectives. To fully exploit the sample similarity relationships in different kernel spaces, OMELET employs multiple kernel learning to learn a consensus kernel and obtain the similarity matrix between samples. To eliminate noise from the similarity matrix captured in the fused kernel space, OMELET leverages topological similarity propagated among highly similar neighbors to reshape the similarity graph. Finally, OMELET captures a topological affinity matrix that represents the sample similarity relationships. Moreover, to enable OMELET to scale on large datasets, we derive a variant called OMELET-SV, which leverages anchors to construct an anchor-based kernel matrix, thereby improving computational efficiency. Extensive experimental results demonstrate the outstanding performance of OMELET compared to 18 representative methods.

Multi-view outlier detection is essential for analyzing complex data, yet many existing approaches rely on shallow fusion schemes or remain sensitive to noise. We propose a Bilateral Graph Filtering Framework with Alternating Optimization for Robust Multi-View Outlier Detection(BGFMOD), which detects attribute outliers via high-frequency residuals and identifies class outliers through cross-view structural inconsistencies. An alternating optimization strategy efficiently addresses the non-convexity of bilateral filtering while preserving spectral orthogonality. Experiments on multiple benchmark datasets show that BGFMOD achieves superior AUC performance and strong robustness across diverse multi-view scenarios. The complete code is available at https://github.com/Jinzhao11/BGFMOD

Dynamic text-attributed graphs (DTAGs) contain time-stamped interactions with evolving textual information. Existing DTAG models, including recent LLM-based approaches, typically perform semantic modeling at uniformly sampled timestamps, despite the fact that semantic changes and structural evolution are often asynchronous. This leads to redundant semantic computation at uninformative times and missed moments of significant semantic shift. We propose TDiscNet, a topology-guided framework that adaptively decides when to perform semantic modeling by measuring the mismatch between semantic and structural evolution. TDiscNet constructs direction-aware semantic graphs from historical texts, computes persistence diagrams under a shared time-based filtration, and defines a Topological Discrepancy Score via the Wasserstein distance between diagrams. The resulting score serves as a criterion for node-wise timestamp sampling, assigning higher priority to timestamps with stronger semantic-structural discrepancy. Experiments on multiple real world DTAG benchmarks demonstrate that TDiscNet improves temporal link prediction while performing semantic modeling on a smaller subset of sampled timestamps, thereby reducing semantic computation compared with uniform schedules.

Computer-Aided Design (CAD) models are widely used across industrial design, simulation, and manufacturing processes. Text-to-CAD systems aim to generate editable, general-purpose CAD models from textual descriptions, significantly reducing the complexity and entry barrier associated with traditional CAD workflows. However, rendering CAD models can be slow, and deploying VLMs to review CAD models can be expensive and may introduce reward hacking that degrades the systems. To address these challenges, we propose CAD-Judge, a novel, verifiable reward system for efficient and effective CAD preference grading and grammatical validation. We adopt the Compiler-as-a-Judge Module (CJM) as a fast, direct reward signal, optimizing model alignment by maximizing generative utility through prospect theory. To further improve the robustness of Text-to-CAD in the testing phase, we introduce a simple yet effective agentic CAD generation approach and adopt the Compiler-as-a-Review Module (CRM), which efficiently verifies the generated CAD models, enabling the system to refine them accordingly. Extensive experiments on challenging CAD datasets demonstrate that our method achieves state-of-the-art performance while maintaining superior efficiency.

Signed networks encode both cooperative and antagonistic relationships, providing a powerful framework for modeling complex relational dynamics. However, most existing embedding methods are grounded in structural balance theory and rely on geometric proximity to preserve link polarity and balanced structures. Consequently, these approaches often fail to capture structurally unbalanced patterns that play a crucial role in bridging polarized communities while exaggerating group-level polarization in the original network. Recent studies reveal that node attribute heterogeneity is closely related to the emergence of such unbalanced patterns. To overcome this limitation, we propose a novel polarity link modeling framework that integrates geometric proximity and attribute similarity to jointly determine both the existence and sign of links. To validate its applicability to real-world signed networks, we instantiate this framework into a generative model, enabling the controlled synthesis of networks with varying proportions of balanced and unbalanced structures. Building on this, we introduce ISNE, an Imbalance-aware Signed Network Embedding method that dynamically adapts to the structural balance ratio of the input network. By maximizing the likelihood function between the observed and generated networks, ISNE learns node representations that not only preserve link polarity and balanced structures but also faithfully capture functionally significant unbalanced structures. Extensive experiments on synthetic and real-world signed networks demonstrate that ISNE substantially outperforms existing methods in link sign prediction tasks, particularly in networks with high proportions of unbalanced structures. These findings highlight the importance of modeling unbalanced patterns and the effectiveness of ISNE in advancing signed network representation.