Dehua Peng

WenZhang Wei, Zhipeng Gui, Dehua Peng et al.

The core of vision-language models lies in measuring cross-modal similarity within a unified representation space. However, most image-text matching or multi-class image classification datasets lack fine-grained cross-modal matching annotations, forcing the continuous similarity space into binary classification boundaries. This compression induces false negative samples and significantly impairs the generalization performance of cross-modal tasks. While prior research has attempted to mitigate this by modeling intra-modal ambiguity, it often overlooks inherent annotation flaws, leading to suboptimal uncertainty allocation. To address these challenges, we propose a Variational Adapter for Cross-modal Similarity Representation (VACSR). This approach reformulates image-text matching with fine-grained semantic scarcity as a variational inference problem. It constructs a latent space for cross-modal similarity and uses regularization techniques to mitigate overfitting to binary annotations. Experiments on image-text retrieval, domain generalization, and base-to-novel generalization demonstrate the proposed method's effectiveness and robust generalization ability.

Wenzhang Wei, Zhipeng Gui, Changguang Wu et al.

The core of cross-modal matching is to accurately measure the similarity between different modalities in a unified representation space. However, compared to textual descriptions of a certain perspective, the visual modality has more semantic variations. So, images are usually associated with multiple textual captions in databases. Although popular symmetric embedding methods have explored numerous modal interaction approaches, they often learn toward increasing the average expression probability of multiple semantic variations within image embeddings. Consequently, information entropy in embeddings is increased, resulting in redundancy and decreased accuracy. In this work, we propose a Dynamic Visual Semantic Sub-Embeddings framework (DVSE) to reduce the information entropy. Specifically, we obtain a set of heterogeneous visual sub-embeddings through dynamic orthogonal constraint loss. To encourage the generated candidate embeddings to capture various semantic variations, we construct a mixed distribution and employ a variance-aware weighting loss to assign different weights to the optimization process. In addition, we develop a Fast Re-ranking strategy (FR) to efficiently evaluate the retrieval results and enhance the performance. We compare the performance with existing set-based method using four image feature encoders and two text feature encoders on three benchmark datasets: MSCOCO, Flickr30K and CUB Captions. We also show the role of different components by ablation studies and perform a sensitivity analysis of the hyperparameters. The qualitative analysis of visualized bidirectional retrieval and attention maps further demonstrates the ability of our method to encode semantic variations.

Dehua Peng, Zhipeng Gui, Jie Gui et al.

Boundary point detection aims to outline the external contour structure of clusters and enhance the inter-cluster discrimination, thus bolstering the performance of the downstream classification and clustering tasks. However, existing boundary point detectors are sensitive to density heterogeneity or cannot identify boundary points in concave structures and high-dimensional manifolds. In this work, we propose a robust and efficient boundary point detection method based on Local Direction Dispersion (LoDD). The core of boundary point detection lies in measuring the difference between boundary points and internal points. It is a common observation that an internal point is surrounded by its neighbors in all directions, while the neighbors of a boundary point tend to be distributed only in a certain directional range. By considering this observation, we adopt density-independent K-Nearest Neighbors (KNN) method to determine neighboring points and design a centrality metric LoDD using the eigenvalues of the covariance matrix to depict the distribution uniformity of KNN. We also develop a grid-structure assumption of data distribution to determine the parameters adaptively. The effectiveness of LoDD is demonstrated on synthetic datasets, real-world benchmarks, and application of training set split for deep learning model and hole detection on point cloud data. The datasets and toolkit are available at: https://github.com/ZPGuiGroupWhu/lodd.

Dehua Peng, Zhipeng Gui, Huayi Wu

As the most typical graph clustering method, spectral clustering is popular and attractive due to the remarkable performance, easy implementation, and strong adaptability. Classical spectral clustering measures the edge weights of graph using pairwise Euclidean-based metric, and solves the optimal graph partition by relaxing the constraints of indicator matrix and performing Laplacian decomposition. However, Euclidean-based similarity might cause skew graph cuts when handling non-spherical data distributions, and the relaxation strategy introduces information loss. Meanwhile, spectral clustering requires specifying the number of clusters, which is hard to determine without enough prior knowledge. In this work, we leverage the path-based similarity to enhance intra-cluster associations, and propose MeanCut as the objective function and greedily optimize it in degree descending order for a nondestructive graph partition. This algorithm enables the identification of arbitrary shaped clusters and is robust to noise. To reduce the computational complexity of similarity calculation, we transform optimal path search into generating the maximum spanning tree (MST), and develop a fast MST (FastMST) algorithm to further improve its time-efficiency. Moreover, we define a density gradient factor (DGF) for separating the weakly connected clusters. The validity of our algorithm is demonstrated by testifying on real-world benchmarks and application of face recognition. The source code of MeanCut is available at https://github.com/ZPGuiGroupWhu/MeanCut-Clustering.

Dehua Peng, Zhipeng Gui, Huayi Wu

The characteristics of data like distribution and heterogeneity, become more complex and counterintuitive as dimensionality increases. This phenomenon is known as curse of dimensionality, where common patterns and relationships (e.g., internal pattern and boundary pattern) that hold in low-dimensional space may be invalid in higher-dimensional space. It leads to a decreasing performance for the regression, classification, or clustering models or algorithms. Curse of dimensionality can be attributed to many causes. In this paper, we first summarize the potential challenges associated with manipulating high-dimensional data, and explains the possible causes for the failure of regression, classification, or clustering tasks. Subsequently, we delve into two major causes of the curse of dimensionality, distance concentration, and manifold effect, by performing theoretical and empirical analyses. The results demonstrate that, as the dimensionality increases, nearest neighbor search (NNS) using three classical distance measurements, Minkowski distance, Chebyshev distance, and cosine distance, becomes meaningless. Meanwhile, the data incorporates more redundant features, and the variance contribution of principal component analysis (PCA) is skewed towards a few dimensions.

Dehua Peng, Zhipeng Gui, Wenzhang Wei et al.

As a pivotal branch of machine learning, manifold learning uncovers the intrinsic low-dimensional structure within complex nonlinear manifolds in high-dimensional space for visualization, classification, clustering, and gaining key insights. Although existing techniques have achieved remarkable successes, they suffer from extensive distortions of cluster structure, which hinders the understanding of underlying patterns. Scalability issues also limit their applicability for handling large-scale data. We hence propose a sampling-based Scalable manifold learning technique that enables Uniform and Discriminative Embedding, namely SUDE, for large-scale and high-dimensional data. It starts by seeking a set of landmarks to construct the low-dimensional skeleton of the entire data, and then incorporates the non-landmarks into the learned space based on the constrained locally linear embedding (CLLE). We empirically validated the effectiveness of SUDE on synthetic datasets and real-world benchmarks, and applied it to analyze single-cell data and detect anomalies in electrocardiogram (ECG) signals. SUDE exhibits distinct advantage in scalability with respect to data size and embedding dimension, and has promising performance in cluster separation, integrity, and global structure preservation. The experiments also demonstrate notable robustness in embedding quality as the sampling rate decreases.

Yuan Wang, Zhipeng Gui, Huayi Wu et al.

With increasing point of interest (POI) datasets available with fine-grained spatial and temporal attributes, space-time Ripley's K function has been regarded as a powerful approach to analyze spatiotemporal point process. However, space-time Ripley's K function is computationally intensive for point-wise distance comparisons, edge correction and simulations for significance testing. Parallel computing technologies like OpenMP, MPI and CUDA have been leveraged to accelerate the K function, and related experiments have demonstrated the substantial acceleration. Nevertheless, previous works have not extended optimization of Ripley's K function from space dimension to space-time dimension. Without sophisticated spatiotemporal query and partitioning mechanisms, extra computational overhead can be problematic. Meanwhile, these researches were limited by the restricted scalability and relative expensive programming cost of parallel frameworks and impeded their applications for large POI dataset and Ripley's K function variations. This paper presents a distributed computing method to accelerate space-time Ripley's K function upon state-of-the-art distributed computing framework Apache Spark, and four strategies are adopted to simplify calculation procedures and accelerate distributed computing respectively. Based on the optimized method, a web-based visual analytics framework prototype has been developed. Experiments prove the feasibility and time efficiency of the proposed method, and also demonstrate its value on promoting applications of space-time Ripley's K function in ecology, geography, sociology, economics, urban transportation and other fields.