Is this you? Claim this profile to correct it, add a bio and choose the work people see first.
Claim this profileAcademic lineage
Possible advisorsa guess from early papers, not confirmed
- Lei XuSuggested from co-authorship
Is this you? Claim this profile to confirm or dismiss it.
Works531 from public data
- Segmentation of retinal blood vessels using the radial projection and semi-supervised approach402
A novel scheme of extracting the retinal vessels based on the radial projection and semi-supervised method that can achieve improved detection of thin vessels and decrease false detection of vessels in pathological regions compared to rival solutions is presented.
- 287
- 234
- -Means: A new generalized k-means clustering algorithm229
A generalized version of the conventional k-means clustering algorithm applicable to ellipse-shaped data clusters without dead-unit problem, but also performs correct clustering without pre-assigning the exact cluster number.
- MTFH: A Matrix Tri-Factorization Hashing Framework for Efficient Cross-Modal Retrieval220
This paper exploits an efficient objective function to flexibly learn the modality-specific hash codes with different length settings, while synchronously learning two semantic correlation matrices to semantically correlate the different hash representations for heterogeneous data comparable, so that the derived hash codes are more semantically meaningful for various challenging cross-modal retrieval tasks.
- 207
- Joint sparse principal component analysis203
A novel approach called joint sparse principal component analysis (JSPCA) is proposed to jointly select useful features and enhance robustness to outliers and the experimental results demonstrate that the proposed approach is feasible and effective.
- Long-tailed Visual Recognition via Gaussian Clouded Logit Adjustment177
This paper proposes the Gaussian clouded logit adjustment by Gaussian perturbation of different class logits with varied amplitude, and proposes the class-based effective number sampling strategy with classifier retraining to alleviate the bias in a classifier.
- 165
- 150
- 145
- Intelligent data engineering and automated learning : 4th International Conference, IDEAL 2003, Hong Kong, China, March 21-23, 2003 : revised papers136
It is shown how one can perform practical simulation by implementing a guiding function to the conventional Monte Carlo method to search for the optimal path of the traveling salesman problem and its performance is comparable with more elaborate and heuristic methods.
- Efficient single image dehazing and denoising: An efficient multi-scale correlated wavelet approach135
This paper presents an efficient multi-scale correlated wavelet approach to solve the image dehazing and denoising problem in the frequency domain and finds a generic regularity in nature images that the haze is typically distributed in the low frequency spectrum of its multi- scale wavelet decomposition.
- 133
- 131
- Categorical-and-numerical-attribute data clustering based on a unified similarity metric without knowing cluster number126
A general clustering framework based on the concept of object-cluster similarity is presented and gives a unified similarity metric which can be simply applied to the data with categorical, numerical, and mixed attributes.
- 124
- Independent component ordering in ICA time series analysis103
This paper proposes an alternative approach to order the components according to their joint contributions in data reconstruction, which naturally leads the component ordering to a typical combinatorial optimization problem, whereby the underlying optimum ordering can be found in an exhaustive way.
- 101
- 100
- 97
- 94
- Bayes Imbalance Impact Index: A Measure of Class Imbalanced Data Set for Classification Problem90
IBI3 can be used as an instance complexity measure of imbalance and BI3 as a criterion to demonstrate the degree to which imbalance deteriorates the classification of a data set and is used to access whether it is worth using imbalance recovery methods to recover the performance loss of a classifier.
- Learning Self-Growth Maps for Fast and Accurate Imbalanced Streaming Data Clustering85
It turns out that the proposed SGM can incrementally adapt to the distributions of new chunks, and the self-growth map-guided hierarchical merging for the imbalanced data clustering (SOHI) approach can quickly explore a true number of imbalanced clusters.
- 84
- Dynamic Weighted Majority for Incremental Learning of Imbalanced Data Streams with Concept Drift84
Experiments on both synthetic and real data sets with concept drift show that DWMIL performs better than the state-of-the-art competitors, with less computational cost.
- 82
- A new feature selection method for Gaussian mixture clustering82
This paper proposes a new feature selection method, through which not only the most relevant features are identified, but the redundant features are also eliminated so that the smallest relevant feature subset can be found.
- Feature Extraction for Incomplete Data Via Low-Rank Tensor Decomposition With Feature Regularization79
- Improving posture classification accuracy for depth sensor-based human activity monitoring in smart environments77
A framework that accurately classifies the nature of the 3D postures obtained by Kinect using a max-margin classifier that consistently outperforms existing algorithms in both activity classification and posture healthiness classification is proposed.
- Reversible data hiding in Paillier cryptosystem76
Two new reversible data hiding algorithms are proposed for the encrypted images that are more applicable in cloud computing without degrading the security level and can be losslessly transmitted in the different applications of Paillier cryptosystem.
- Feature Fusion from Head to Tail for Long-Tailed Visual Recognition75
This work proposes to augment tail classes by grafting the diverse semantic information from head classes, referred to as head-to-tail fusion (H2T), which replaces a portion of feature maps from tail classes with those belonging to head classes, substantially enhance the diversity of tail classes.
- ApLeaf: An efficient android-based plant leaf identification system73
An Android-based mobile application designed to automatically identify plant species according to the photographs of tree leaves is described and works well with state-of-the-art identification performance.
- FDDH: Fast Discriminative Discrete Hashing for Large-Scale Cross-Modal Retrieval71
An efficient fast discriminative discrete hashing approach is proposed for large-scale cross-modal retrieval and an effective and stable online learning strategy is presented for optimizing modality-specific projection functions, featuring adaptivity to different training sizes and streaming data.
- 71
- Sparse discriminative multi-manifold embedding for one-sample face identification71
This paper proposes a two-step scheme for the OSPP problem by posing it as a representation and matching problem, and presents a novel manifold embedding algorithm, namely sparse discriminative multi-manifold embedding (SDMME), to learn the intrinsic representation beneath the raw data.
- Selected Papers from the Ninth International Conference on Computational Intelligence and Security71
The 2012 International Conference on Computational Intelligence and Security provides a platform to explore the potential applications of CI models, algorithms, and technologies to IS.
- 69
- 68
- 67
- On rival penalization controlled competitive learning for clustering with automatic cluster number selection67
This paper investigates the RPCL and proposes the rival penalization controlled competitive learning algorithm and its stochastic version and compares the performance of RPCCL to RPCL in Gaussian mixture clustering and color image segmentation.
- 66
- 66
- Long-Tailed Visual Recognition via Self-Heterogeneous Integration with Knowledge Excavation65
This work proposes a novel MoE-based method called Self-Heterogeneous Integration with Knowledge Excavation (SHIKE), and proposes Depth-wise Knowledge Fusion (DKF) to fuse features between different shallow parts and the deep part in one network for each expert, which makes experts more diverse in terms of representation.
- Classifying protein sequences using hydropathy blocks64
A new method for classifying protein sequences based upon the hydropathy blocks occurring in protein sequences by utilizing the support vector machine (SVM) classifier to classify the protein sequences into the known protein families.
- Interpreting and Improving Large Language Models in Arithmetic Calculation63
This work serves as a preliminary exploration into the arithmetic calculation abilities inherent in LLMs, laying a solid foundation to reveal more intricate mathematical tasks.
- 59
- FEDIC: Federated Learning on Non-IID and Long-Tailed Data via Calibrated Distillation59
This paper studies the joint problem of non-IID and long-tailed data in federated learning and proposes a corresponding solution called Federated Ensemble Distillation with Imbalance Calibration (FEDIC), a new distillation method with logit adjustment and calibration gating network to solve the long-tail problem effectively.
- 59
- 58
- 57
- 56
- 56
- 56
- 55
- 55
- 55
- 54
- 54
- 53
- 52
- Cross-Modal Hashing Method With Properties of Hamming Space: A New Perspective51
A novel algorithm, namely Semantic Channel Hashing (SCH), which classify sample pairs into fully semantic-similar, partially semantic-similar, and semantic-negative ones based on their similarity and impose different constraints on them, respectively to ensure that the entire Hamming space is utilized.
- A high-capacity reversible data hiding method for homomorphic encrypted images51
A new method is proposed for homomorphic encrypted images so that part of the hidden data can be extracted in encrypted domain and the rest are extractable after image decryption.
- Writer identification using fractal dimension of wavelet subbands in gabor domain51
This paper proposes one new method for off-line, text-independent writer identification by using the fractal dimension of wavelet subbands in Gabor domain of the handwriting images to extract more effective features to distinguish the handwritings, and hence achieve much better identification results.
- A fragile watermarking scheme for 3D meshes51
A new fragile watermarking scheme for 3D meshes that is invariant to translation, rotation and uniformly scaling, but sensitive to other operations, and the embedding strength is adjustable so that a trivial tampering with the watermarked mesh model may lead the embedded watermark to change.
- Single object tracking via robust combination of particle filter and sparse representation50
This paper proposes a tracking method based on the robust combination of particle filter and reverse sparse representation (RC-PFRSR) to reduce the drifting and proposes a diagonal matrix α, whose diagonal line includes each patch contribution factor, to function each patch coefficient value of one candidate obtained by sparse representation.
- 49
- 49
- Contrastive Learning Assisted-Alignment for Partial Domain Adaptation48
A contrastive learning-assisted alignment (CLA) approach for PDA to jointly align distributions across domains for better adaptation and to reweight source instances to reduce the contribution of outlier instances is proposed.
- 48
- 48
- Graph-Based Dissimilarity Measurement for Cluster Analysis of Any-Type-Attributed Data47
A new dissimilarity metric for cluster analysis of heterogeneous attribute data, which computes the dissimilarities between attribute values under the guidance of the graph structures and develops a new clustering algorithm associated with this proposed metric.
- Learnable Weighting of Intra-attribute Distances for Categorical Data Clustering with Nominal and Ordinal Attributes47
This paper proposes a novel distance metric to measure the intra-attribute distances of nominal and ordinal attributes in a unified way, meanwhile preserving the order relationship among ordinal values and proposes a new clustering algorithm to make the learning of intra- attribute distance weights and partitions of data objects into a single learning paradigm rather than two separate steps, whereby circumventing a suboptimal solution.
- Reversible Data Hiding With Image Contrast Enhancement Based on Two-Dimensional Histogram Modification47
The method based on the 2D histogram modification is proposed to perform image CE and RDH simultaneously and is given to show its superior performances in CE and image quality preservation by comparing with the state-of-the-art approaches.
- Robust heterogeneous discriminative analysis for face recognition with single sample per person47
This work presents a new graph-based Fisher-like criterion, which incorporates two manifold embeddings, to learn heterogeneous discriminative representations of image patches, and proposes a new patch-based method called Robust Heterogeneous Discriminative Analysis (RHDA), for FR with SSPP.
- 47
- A novel approach to extracting features from motif content and protein composition for protein sequence classification47
This paper presents a novel approach to extracting features from motif content and protein composition for protein sequence classification by utilizing the SVM classifiers to classify protein sequences into corresponding families based on the selected feature subsets.
- 45
- Multi-Class Imbalance Classification Based on Data Distribution and Adaptive Weights44
This work incorporates the class imbalance ratio, the density-based factor, and the adaptive weighting mechanism into a single variable, based on which the adaptive weights of all training samples are computed, to further alleviate the multi-class imbalance issue.
- 44
- 43
- 43
- Label-Noise Learning with Intrinsically Long-Tailed Data42
This paper proposes two-stage bi-dimensional sample selection (TABASCO), a learning frame-work for label-noise learning with intrinsically long-tailed data to better separate clean samples from noisy samples, especially for the tail classes.
- 41
- 41
- A Multi-Objective Evolutionary Algorithm Using Min-Max Strategy And Sphere Coordinate Transformation41
The proposed algorithm can overcome the drawbacks of the algorithms using the weighted sum of the objectives, and explore the objective space to approximate uniformly distributed solutions on the Pareto front gradually.
- 40
- 40
- 39
- 39
- Learning a mixture model for clustering with the completed likelihood minimum message length criterion39
This paper tackles the model selection and parameter estimation problems in model-based clustering so as to improve the clustering performance on the data sets whose true kernel distribution functions are not in the family of assumed ones, as well as with inherently overlapped clusters.
- Security of public key encryption technique based on multiple chaotic systems39
Both theoretical and experimental results show that the attacker can access to the secret key without difficulty and the lack of security discourages the use of such algorithm for practical applications.
- Triplet Fusion Network Hashing for Unpaired Cross-Modal Retrieval38
A one-stream framework triplet fusion network hashing (TFNH) is proposed, which mainly consists of two parts, which are used to utilize both paired and unpaired data and embed manifold learning into the framework for preserving both inter and intra modal similarity.
- Tensor Rank Estimation and Completion via CP-based Nuclear Norm37
Tensor Rank Estimation based on $L_1$-regularized orthogonal CP decomposition (TREL1) is proposed, which incorporates a regularization with CP-based tensor nuclear norm when minimizing the reconstruction error in TC to automatically determine the rank of an incomplete tensor.
- 37
- Learning to Shape In-distribution Feature Space for Out-of-distribution Detection36
This paper gives an affirmative answer to this question by presenting a Distributional Representation Learning ( DRL ) framework for OOD detection, in particular, DRL explicitly enforces the underlying feature space to conform to a pre-defined mixture distribution, together with an online approximation of normalization constants to enable end-to-end training.
- 36
- 35
- 35
- 34
- 33
- 33
- Hybrid Sampling with Bagging for Class Imbalance Learning33
Experiments show that the proposed hybrid sampling method significantly outperforms the other state-of-the-art bagging-based methods for class imbalance problem and the superiority of sampling rate selection is also demonstrated.
- 33
- 33
- 33
- Automatic lip localization under face illumination with shadow consideration32
This paper proposes a new approach to obtain the minimum enclosing rectangle surround of a mouth automatically based upon the transformed gray-level image and shows the promising result of the proposed approach in comparison with the existing methods.
- 31
- A local region based approach to lip tracking31
A local region based approach to lip tracking, which consists of two phases: lip contour extraction for the first lip frame, and followed by lip tracking in the subsequent lip frames, which is efficient in comparison with the existing methods.
- Cross-Graph Attention Enhanced Multi-Modal Correlation Learning for Fine-Grained Image-Text Retrieval30
This work proposes a Cross-Graph Attention model (CGAM) to explicitly learn the shared semantic concepts, which can be well utilized to guide the feature learning process of each modality and promote the common embedding learning.
- An Ordinal Data Clustering Algorithm with Automated Distance Learning30
This paper proposes a novel ordinal data clustering algorithm, which iteratively learns: 1) The partition of ordinal dataset, and 2) the inter-category distances, and is parameter-free.
- Inference-Based Similarity Search in Randomized Montgomery Domains for Privacy-Preserving Biometric Identification30
This paper proposes an inference-based framework for privacy-preserving similarity search in Hamming space that builds on an obfuscated distance measure that can conceal Hamming distance in a dynamic interval and proposes to apply Montgomery multiplication for generating search indexes that can withstand adversarial similarity analysis.
- Corrupted and occluded face recognition via cooperative sparse representation30
A CoSR model is introduced, which combines the fSR and the bSR together, into robust face recognition, by proposing two alternative methods to these two key problems: learning bases and unknown faces help to enrich the bases set of the b SR.
- Automatic motion capture data denoising via filtered subspace clustering and low rank matrix approximation30
An automatic Motion Capture data denoising approach via filtered subspace clustering and low rank matrix approximation that does not need any physical information about the underling structure of MoCap data or require auxiliary data sets for training priors is presented.
- Cooperative and penalized competitive learning with application to kernel-based clustering30
A novel learning algorithm called Cooperative and Penalized Competitive Learning (CPCL), which implements the cooperation and penalization mechanisms simultaneously in a single competitive learning process, to locate the cluster centers more quickly and be insensitive to the number of seed points and their initial positions.
- 29
- 29
- 29
- A Data Mapping Method for Steganography and Its Application to Images29
A new steganographic method that preserves the first-order statistics of the cover that is suitable for the passive warden scenario, but not robust to any change of the stego object is proposed.
- Learning unified distance metric for heterogeneous attribute data clustering28
This paper proposes a novel Heterogeneous Attribute Reconstruction and Representation (HARR) learning paradigm accordingly for cluster analysis that transforms heterogeneous attributes into a homogeneous status for distance metric learning, and integrates the learning with clustering to automatically adapt the metric to different clustering tasks.
- Variational Multi-scale Representation for Estimating Uncertainty in 3D Gaussian Splatting28
By leveraging the model parameter uncertainty estimated by the proposed uncertainty estimation method, the method can remove noisy Gaussians automatically, thereby obtaining a high-fidelity part of the reconstructed scene, which is of great help in improving the visual quality.
- 28
- 28
- Arabic font recognition based on diacritics features28
This paper presents the diacritics as the thumb of Arabic fonts which can be used individually to identify and recognize the font type and it can be integrated with OCR systems very easily.
- A Novel Hybrid GA/SVM System for Protein Sequences Classification28
A novel hybrid genetic algorithm(GA)/Support Vector Machine (SVM) system, which selects features from the protein sequences and trains the SVM classifier simultaneously using a multi-objective genetic algorithm, is proposed in this paper.
- Iterative Dynamic Generic Learning for Face Recognition From a Contaminated Single-Sample Per Person27
- High-Capacity Reversible Data Hiding in Encrypted Images by Bit Plane Partition and MSB Prediction27
The experimental results show that higher embedding capacity can be achieved with the proposed RDH scheme, compared with the state-of-the-art RDH schemes for encrypted images.
- QGRL: Quaternion Graph Representation Learning for Heterogeneous Feature Data Clustering26
Extensive experimental evaluations illustrate the superiority of the proposed Quaternion Graph Representation Learning (QGRL) method in terms of clustering accuracy and robustness to various data sets composed of arbitrary combinations of numerical and categorical features.
- 26
- 26
- 26
- 26
- 25
- Het2Hom: Representation of Heterogeneous Attributes into Homogeneous Concept Spaces for Categorical-and-Numerical-Attribute Data Clustering25
A novel attribute representation learning method called Het2Hom is proposed, which first converts the heterogeneous attributes into a homogeneous form, and then learns attribute representations and data partitions on such a homogeneity basis.
- 25
- Solving Dynamic Multi-objective Optimization Problems Using Incremental Support Vector Machine25
This paper incorporates the proposed approach to reusing the obtained Pareto optimal set to train prediction models via machine learning approaches into three evolutionary algorithms, the multi-objective particle swarm optimization (MOPSO), Nondominated Sorting Genetic Algorithm II (NSGA-II), and the Regularity Model-based multi- objective estimation of distribution algorithm (RE-MEDA).
- 25
- Plant Leaf Identification via a Growing Convolution Neural Network with Progressive Sample Learning25
Experiments and analyses show that the proposed GCNN outperforms other state-of-the-art algorithms such as the traditional CNN and the hand-crafted features with SVM classifiers.
- Topology Aware Data‐Driven Inverse Kinematics25
A new data-driven inverse kinematics framework that conserves the topology of the synthesizing postures using the Gauss Linking Integral (GLI), such that penetration can be efficiently prevented in complex movements.
- 25
- A competitive and cooperative learning approach to robust data clustering.25
A new semi-competitive learning paradigm named Competitive and Cooperative Learning (CCL), in which seed points not only compete each other for updating to adapt to an input each time, but also dynamically cooperate to achieve the learning task.
- An extended asld trading system to enhance portfolio management25
An extension of the adaptive supervised learning decision (ASLD) system (EASLD), which combines the ASLD with a portfolio optimization scheme to take a balance between the expected returns and risks.
- 24
- 24
- Learning Discriminative Joint Embeddings for Efficient Face and Voice Association24
This paper proposes to learn discriminative joint embedding for face-voice association, which can seamlessly train the face subnetwork and voice subnetwork to learn their high-level semantic features, while correlating them to be compared directly and efficiently.
- Fast hypervolume approximation scheme based on a segmentation strategy24
A fast hypervolume approximation method with both improved speed and accuracy than the previous approximation methods via a new segmentation strategy is developed, and the performance of the integrated algorithm is also very competitive among the experimental algorithms.
- 24
- Optimal WCDMA network planning by multiobjective evolutionary algorithm with problem-specific genetic operation24
The proposed encoding method and genetic operation are merged into the multiobjective evolutionary algorithm-based decomposition (MOEA/D-M2M) to solve the WCDMA network planning problem.
- 23
- Joint Semantic Preserving Sparse Hashing for Cross-Modal Retrieval21
A new concept of cluster-wise semantic relationship is introduced, which leverages label distribution to indicate which samples are more suitable for clustering and preserves both kinds of semantic relationships to ensure that more samples with similar semantics are clustered together, thereby achieving better retrieval results.
- Feature-Balanced Loss for Long-Tailed Visual Recognition21
This paper addresses the long-tailed problem from feature space and proposes the feature-balanced loss, which encourages larger feature norms of tail classes by giving them relatively stronger stimuli.
- Adaptive micro partition and hierarchical merging for accurate mixed data clustering20
It turns out that the proposed approach breaks through the clustering performance bottleneck brought by the pre-set number of sought clusters k and cluster distribution bias, and is thus capable of clustering datasets comprising various combinations of numerical and categorical attributes.
- Adjusting Logit in Gaussian Form for Long-Tailed Visual Recognition20
This article introduces feature augmentation to balance the embedding distribution of long-tailed visual recognition from the perspective of feature level and proposes two novel logit adjustment methods to improve model performance at a modest computational overhead.
- Realize Generative Yet Complete Latent Representation for Incomplete Multi-View Learning20
A novel variational lower bound is introduced to integrate view-invariant information into posterior inference to enhance the solidative of the learned latent representation of Complete Multi-view Variational Auto-Encoders.
- Hearing like Seeing20
A novel adversarial deep semantic matching network for efficient voice-face interactions and associations is presented, which can well learn the correspondence between voices and faces for various cross-modal matching and retrieval tasks.
- 20
- Attention guided deep audio-face fusion for efficient speaker naming20
This paper treats the speaker naming task as a group of matched audio-face pair finding problems, and presents an efficient attention guided deep audio- face fusion approach to detect the active speakers and shows its very competitive performance with the state of thearts.
- 20
- A High-Capacity Data Hiding Method for Polygonal Meshes20
A secret key is used to order the process of mesh traversal so that it is even harder to construct a counterfeit mesh with the same watermark, and the security is improved by making it hard to estimate the quantization step from the modified distances.
- 20
- 19
- 19
- Reversible contrast enhancement for medical images with background segmentation19
The experimental results on 30 chest radiograph images and 20 magnetic resonance images have demonstrated the efficacy of the proposed scheme for reversible CE and the better performances are provided in achieving CE effects and preserving image quality.
- Lossless Contrast Enhancement of Color Images with Reversible Data Hiding19
The evaluation results demonstrate that better image quality and increased information entropy can be simultaneously achieved with the proposed HSV (hue, saturation, and value) color model.
- Pseudo-random number generator based on asymptotic deterministic randomness19
Both theoretical and experimental results show that the symbolic sequence of the asymptotic deterministic randomness possesses very good cryptographic properties, which improve the security of chaos based PRBGs and increase the resistance against entropy attacks and symbolic dynamics attacks.
- 19
- 18
- 18
- 18
- 18
- 18
- 18
- A Unified Metric for Categorical and Numerical Attributes in Data Clustering18
This paper presents a general clustering framework based on the concept of object-cluster similarity and gives a unified similarity metric which can be simply applied to the data with categorical, numerical, or mixed attributes.
- 18
- 18
- Multiview Tensor Spectral Clustering via Co-Regularization17
This work proposes to achieve multi-view uniform clustering via consensus representation co-regularization through aligning the multi-view low-dimensional representation by co-regularization, and demonstrates the effectiveness of this method in comparison with the recent thirteen benchmark methods.
- 17
- 17
- 17
- 17
- 17
- 17
- Automatic Segmentation of Color Lip Images Based on Morphological Filter17
A new segmentation approach to lip contour extraction is presented by taking account of the color difference between lip and skin in color space by establishing a Gaussian model and making use of the hue and saturation value of each pixel within the lip segment to estimate the model parameters.
- 17
- 16
- UGNCL: Uncertainty-Guided Noisy Correspondence Learning for Efficient Cross-Modal Matching16
This work proposes an efficient Uncertainty-Guided Noisy Correspondence Learning (UGNCL) framework, which is reliably designed to divide the data into clean, noisy and hard partitions, which can effortlessly mitigate the impact of easily-determined noisy pairs.
- Facial Structure Guided GAN for Identity-preserved Face Image De-occlusion16
This paper addresses the problem of removing occlusions on face images and presents a new two-stage Facial Structure Guided Generative Adversarial Network (FSG-GAN), aiming at further eliminating the appearance ambiguity as well as unnatural texture.
- A branch and bound irredundant graph algorithm for large-scale MLCS problems16
A branch and bound irredundant graph algorithm called Big-MLCS is proposed, which constructs a much smaller DAG than the existing algorithms do by a Branch and bound method, and designs a new data structure to efficiently store and handle Small-DAG.
- 16
- 16
- 16
- 16
- 15
- 15
- Server-Client Collaborative Distillation for Federated Reinforcement Learning15
This article is among the first attempts to analyze the heterogeneity problem in FRL and propose an off-policy FRL framework, termed as Server-Client Collaborative Distillation (SCCD), which has a superior performance in four classical continuous control tasks with non-IID environments.
- 15
- 15
- Variance constrained partial least squares15
This paper derived the mathematical formulation of the relationship between mean square error (MSE) and these three factors and found that MSE is not monotonous with the product of the three factors.
- 15
- 15
- 14
- Improve Knowledge Distillation via Label Revision and Data Selection14
This article proposes to rectify the teacher's inaccurate predictions using the ground truth to rectify the incorrect supervision and introduces a data selection technique to choose suitable training samples to be supervised by the teacher, thereby reducing the impact of incorrect predictions to some extent.
- 14
- Proximal average approximated incremental gradient descent for composite penalty regularized empirical risk minimization14
A new PA-based algorithm called IncrePA is proposed by incorporating PA approximation into an incremental gradient framework that features lower per-iteration cost, a faster convergence rate for convex Composite penalties, and guaranteed convergence for even nonconvex composite penalties.
- 14
- Topological fractal networks introduced by mixed degree distribution14
This work investigates networks growth mechanism of the branching-deactivated geographical attachment preference that learned from certain empirical evidence of social behaviors, and finds perfect topological fractal structure emerges by a mixture of both algebraic and exponential degree distributions in a wide range of parameter values.
- Adaptive Rival Penalized Competitive Learning and Combined Linear Predictor Model for Financial Forecast and Investment14
Experiments have shown that Adaptive RPCL-CLP not only outperforms the other approaches with the smallest prediction error and training costs, but also brings in considerable high profits in the trading simulation of foreign exchange market.
- Trending Applications of Large Language Models: A User Perspective Survey13
This survey is positioned as a gateway to LLM literacy, aims to bridges the gap between the cutting-edge LLMs and potential users, and demystifies LLM capabilities through intuitive taxonomies, case studies, and hands-on evaluations.
- Component-Level Oracle Bone Inscription Retrieval13
A component-level OBI retrieval task, i.e., using an OBI component to retrieve all OBI characters containing this component, and proposes a dual-stream attention-based model and two types of triplets based on components and characters as anchors to model the relationships between components and characters.
- 13
- Compact Neural Network via Stacking Hybrid Units13
A new method termed BUnit-Net is presented, which directly constructs compact NNs by stacking designed basic units, without requiring additional judgement criteria anymore, and provides a new promising way for network compression.
- Vertical Federated Principal Component Analysis and Its Kernel Extension on Feature-wise Distributed Data13
The federated principal component analysis for vertically partitioned dataset (VFedPCA) method, which reduces the dimensionality across the joint datasets over all the clients and extracts the principal component feature information for downstream data analysis is proposed.
- Content and Privacy Protection in JPEG Images by Reversible Visual Transformation13
A reversible visual transformation algorithm is proposed to protect content in JPEG images by embedding the location information of the selected blocks in a transformed image, so that the original image can be exactly recovered when needed.
- Exploiting Order Information Embedded in Ordered Categories for Ordinal Data Clustering13
This work proposes a specially designed distance metric, which can exploit the order information embedded in the ordered categories for distance measurement and quantifies the distance between two ordinal categories by accumulating the sub-entropies of all the categories ordered between them.
- Multi-view Manifold Learning for Media Interestingness Prediction13
A novel algorithm dubbed Multi-view Manifold Learning (M) is designed to uncover the latent factors that are capable of distinguishing interesting media data from non-interesting ones by modelling both geometry preserving criterion and discrimination maximization criterion in a unified framework.
- SKM-SNP: SNP markers detection method13
Results indicate that the SKM-SNP method can find some relevant SNPs that categorize the disease samples, and a subspace categorical clustering algorithm is employed to compute a weight for each SNP in the group of patient samples and the groups of normal samples, to identify the subsets of relevantSNPs thatategorize these two groups.
- Feature Selection for Clustering on High Dimensional Data13
A novel feature weighting scheme for a kernel based clustering criterion, in which the weight for each feature is a measure of its contribution to the clustering task, is proposed.
- 13
- 13
- Estimating initial conditions in coupled map lattices from noisy time series using symbolic vector dynamics13
Both theoretical and experimental results show that this algorithm enables us to recover initial condition of CMLs exactly in both noisy and noise free cases, and provides novel analytical techniques for understanding turbulences in coupled map lattices.
- 13
- Preference Bi-objective Evolutionary Algorithm for Constrained Optimization13
A new constraint handling approach that transforms constrained optimization problem of any number of constraints into a two objective preference optimization problem and designs a new crossover operator, a new mutation operator using local search and preference, and a new selection operator based on the preference of the two objectives.
- A new approach to blind source separation with global optimal property.13
It is mathematically proved that the metric value of any linear combination of source signals is less than the largest one of sources under a loose condition and the global optimization of this new metric is achieved by formulating it as a generalized eigenvalue problem.
- PRO-VPT: Distribution-Adaptive Visual Prompt Tuning via Prompt Relocation12
This paper proposes a new VPT framework, termed PRO-VPT (iterative Prompt RelOcation-based VPT), which adaptively adjusts the distribution built upon a nested optimization formulation, and develops a prompt relocation strategy derived from this formulation.
- Lossless Data Hiding in NTRU Cryptosystem by Polynomial Encoding and Modulation12
Two LDH-CT algorithms named Polynomial Encoding (PE) and Polynomial Modulation (PM) are proposed for the “N-th Degree Truncated Polynomial Ring Unit” (NTRU) scheme, respectively and are implemented and compared with several schemes developed for NTRU, BGN, LWE and Paillier encryption.
- Automatic mitral valve leaflet tracking in Echocardiography via constrained outlier pursuit and region-scalable active contours12
This paper presents an automatic mitral leaflet motion tracking approach, which consists of two phases: constrained outlier pursuit for mitral Leaflet detection and its shape refinement and a region-scalable active contour to refine the shapes of the detected mitrals leaflet for final tracking.
- 12
- The asymptotic deterministic randomness12
It is proved that the well-known autonomous system that has been stated to generate deterministic randomness can only act as an approximative asymptotic realizable model and any realizable models for deterministicrandomness will degenerate to some special high dimensional chaotic system.
- A Novel Hybrid GA/RBFNN Technique for Protein Sequences Classification12
Experimental results show that the proposed hybrid GA/RBFNN system outperforms the BLAST and the HMMer.
- 11
- 11
- 11
- Uplink Communication Efficient Differentially Private Sparse Optimization With Feature-Wise Distributed Data11
This paper provides a much generalized convergence analysis for block-coordinate Frank-Wolfe under arbitrary sampling, which greatly extends known convergence results that are only applicable to two specific block sampling distributions.
- Discriminant Manifold Learning via Sparse Coding for Robust Feature Extraction11
A new subspace learning method, namely, discriminant manifold learning via sparse coding (DML_SC) for robust feature extraction, which first decompose each input image into several components via dictionary learning, and then regroup the components into a more important part (MIP) and a less important parts (LIP).
- 11
- 11
- Feature Selection for Local Learning Based Clustering11
This paper assigns a weight to each feature, and incorporates it into the built-in regularization of LLC algorithm to take into account of the relevance of each feature for the clustering, and shows that the resulting weighted regularization with an additional constraint on the weights is equivalent to a known sparse-promoting penalty, thus the weights for irrelevant features can be driven towards zero.
- 11
- An RPCL-based approach for Markov model identification with unknown state number11
This paper presents an alternative identification approach for the Markov model studied in Krishnamurthy and Moore (1993) with the help of a clustering analysis by the rival penalized competitive learning (RPCL) algorithm.
- Deep Tensor Spectral Clustering Network via Ensemble of Multiple Affinity Tensors10
This paper proposes a tensor spectral clustering network (TSC-Net) to achieve one-stage learning of a consensus tensor spectral embedding, while reducing the memory cost.
- 10
- 10
- 10
- 10
- 10
- Improving clustering with pairwise constraints: a discriminative approach10
This paper puts forward a discriminative learning approach which can incorporate pairwise constraints into the recently proposed two-class maximum margin clustering framework, and proposes a set of pairwise loss functions, which features robust detection and penalization for violating the couplewise constraints.
- Lip segmentation and tracking under MAP-MRF framework with unknown segment number10
This paper proposes a lip segmentation and tracking scheme, featuring the robust performance to the estimate of the number of segment clusters, and presents a rival penalized iterative algorithm capable of performing segment clusters and over-segmentation elimination simultaneously.
- 10
- A New Modified Hybrid Learning Algorithm for Feedforward Neural Networks10
A new modified hybrid learning algorithm for feedforward neural networks is proposed to obtain better generalization performance by penalizing both the input-to-output mapping sensitivity and the high frequency components in training data.
- 10
- Letter to the Editor: RIVAL PENALIZED COMPETITIVE LEARNING BASED APPROACH FOR DISCRETE-VALUED SOURCE SEPARATION10
This paper presents an approach based on Rival Penalized Competitive Learning (RPCL) rules for discrete-valued source separation that not only quickly and automatically determines the number of sources, but also is insensitive to the noise in performing blind source separation.
- Shaping pre-trained language models for task-specific embedding generation via consistency calibration9
EGO-PLM is proposed, where PLMs serve as task-specific embedding generator and consistency calibration (CoCa) identifies inconsistencies between the pre-defined and task-specific objectives in an adversarial manner, subsequently calibrating these disparities through adversarial training.
- 9
- Brain Network Classification for Accurate Detection of Alzheimer’s Disease via Manifold Harmonic Discriminant Analysis9
A network manifold harmonic discriminant analysis (MHDA) method is proposed to identify the group-dependent common harmonic waves, which can be used as group-specific references for downstream analyses in Alzheimer’s disease diagnosis with high heterogeneity.
- Password Guessing Based on GAN with Gumbel-Softmax9
Experimental results show the proposed G-Pass outperforms PassGAN in password quality and cracking rate, and by dynamically adjusting one parameter during the training process, a trade-off between sample diversity and quality can be achieved with the proposed model.
- 9
- Feature Extraction for Incomplete Data via Low-rank Tucker Decomposition9
This work proposes an unsupervised method, TDVM, which incorporates low-rank T ucker D ecomposition with feature V ariance M aximization in a unified framework and imposes nuclear norm regularization on the core tensors while minimizing reconstruction errors, and meanwhile maximize the variance of core Tensors.
- Camera Calibration Based on the Common Self-polar Triangle of Sphere Images9
A new linear calibration method is proposed by using the common self-polar triangle of sphere images and recovers five intrinsic parameters without calculating the projected circle center, which is more intuitive and simpler than the previous linear ones.
- 9
- ANALYSIS OF GENE EXPRESSION DATA USING RPEM ALGORITHM IN NORMAL MIXTURE MODEL WITH DYNAMIC ADJUSTMENT OF LEARNING RATE9
This paper applies a one-step approach, namely Rival Penalized Expectation-Maximization (RPEM) algorithm, to analyze the gene expression data and speeds up the learning procedure of RPEM by proposing a new mechanism to adjust the learning rate dynamically.
- 9
- 9
- 8
- 8
- 8
- 8
- Classification of SD-OCT volumes for DME detection: an anomaly detection approach8
The proposed method is based on anomaly detection using Gaussian Mixture Model (GMM) and achieves better classification performances than other recently published works.
- 8
- 8
- Kernel Learning for Local Learning Based Clustering8
A novel kernel learning algorithm within the framework of the Local Learning based Clustering (LLC) (Wu & Scholkopf 2006) that naturally renders a linear combination of kernels.
- SMART: a subspace clustering algorithm that automatically identifies the appropriate number of clusters8
A new penalty term is introduced to the objective function of the fuzzy k-means clustering process to enable several clusters to compete for objects, which leads to merging some cluster centres and the identification of the 'true' number of clusters.
- 8
- 8
- 8
- 8
- k*-Means — A Generalized k-Means Clustering Algorithm with Unknown Cluster Number8
This paper presents a new clustering technique named STepwise Automatic Rival-penalized (STAR) k-means algorithm, which is actually a generalized version of the conventional k-Means (MacQueen 1967), which can perform appropriate clustering without knowing cluster number by gradually penalizing the winning chance of those extra seed points during learning competition.
- ReCon: Enhancing True Correspondence Discrimination through Relation Consistency for Robust Noisy Correspondence Learning7
ReCon leverages a novel relation consistency learning to ensure the dual-alignment of the cross-modal relation consistency between different modalities and the intra-modal relation consistency within modalities, which significantly enhances its effectiveness for true correspondence discrimination.
- Component-Level Segmentation for Oracle Bone Inscription Decipherment7
This paper proposes a new model that leverages a small amount of annotated data and a large amount of weakly annotated data and incorporates expert-provided prior knowledge to automatically segment OBI components, and demonstrates that the method can accurately and clearly present the segmented components to experts.
- 7
- 7
- 7
- GradToken: Decoupling tokens with class-aware gradient for visual explanation of Transformer network7
A gradient-decoupling-based token relevance method (i.e., GradToken) is proposed for the visual explanation of Transformer's predictions, which exploits the class-aware gradient to decouple the tangled semantics in the class token to the semantics corresponding to each category.
- Ask, Attend, Attack: A Effective Decision-Based Black-Box Targeted Attack for Image-to-Text Models7
A decision-based black-box targeted attack as a large-scale optimization problem that achieves targeted attacks without semantic loss on transformer-based and CNN+RNN-based image-to-text models.
- 7
- 7
- 7
- A Fast Objective Reduction Algorithm Based on Dominance Structure for Many Objective Optimization7
This paper presents a novel measure for measuring the capacity of preserving the dominance structure of an objective set, i.e., the redundancy of an Objective to an Objective set, and proposes a fast algorithm to find a minimum set of objectives preserving the dominate structure as much as possible.
- Scalable Spectral k-Support Norm Regularization for Robust Low Rank Subspace Learning7
A scalable and efficient algorithm is proposed which considers the dual objective of the original problem that can take advantage of the more computational efficient linear oracle of the spectral k-support norm to be evaluated and a line-search strategy is adopted in the algorithm to enable it to adapt to the Holder smoothness.
- Efficient generalized conditional gradient with gradient sliding for composite optimization7
A novel algorithm that requires optimal graduate evaluations as proximal gradient for a type of gauge regularized problem where approximation techniques are allowed to further accelerate linear subproblem computation is proposed.
- A Batch Rival Penalized Expectation-Maximization Algorithm for Gaussian Mixture Clustering with Automatic Model Selection7
A batch Rival Penalized Expectation-Maximization (RPEM) algorithm for density mixture clustering provided that all observations are available before the learning process, which is faster than the EM and the adaptive RPEM in general.
- 7
- 7
- 6
- 6
- 6
- 6
- 6
- Reversible Data Hiding in Homomorphic Encrypted Images Without Preprocessing6
A new RDH-EI method for homomorphic encrypted images is proposed by utilizing the additive homomorphism and self-blinding property in Paillier cryptosystem and the experimental results validate that the proposed method is compatible with homomorphic image processing before data embedding.
- 6
- 6
- 6
- 6
- Lip event detection using oriented histograms of regional optical flow and low rank affinity pursuit6
An efficient approach to detecting the visual silence event via the small flow magnitude and a low rank affinity pursuit method to determine the visual speech event that incorporates the lip-dynamic states of mouth opening and closing are addressed.
- Supervised Manifold Learning for Media Interestingness Prediction.6
The models designed for automatically selecting multimedia data, e.g., image and video segments, which are considered to be interesting for a common viewer are described, and a new dimensionality reduction method dubbed Supervised Manifold Regression is introduced to learn the compact representations for predicting the continuous interestingness levels.
- A Novel Evolutionary Multi-objective Algorithm Based on S Metric Selection and M2M Population Decomposition6
A novel S metric selection evolutionary algorithm based on the population decomposition strategy MOEA/D-M2M is proposed to give a simple but effective method to improve the effectiveness of SMS based algorithm.
- 6
- On blind source separation using generalized eigenvalues with a new metric6
It is proved that the metric value of any linear combination of source signals is less than the largest one of sources under a loose condition.
- 6
- 6
- 5
- Mitigating Forgetting in Adapting Pre-trained Language Models to Text Processing Tasks via Consistency Alignment5
A novel dual-model framework, termed as consistency alignment (CoAi), that simulates the pre-training process by performing distributional exploration in the parameter space, and co-training the task-specific model and the auxiliary model to balance the pre-training knowledge and task-specific knowledge during fine-tuning.
- 5
- 5
- Broad Multitask Learning System With Group Sparse Regularization5
An innovative MTL framework explicitly designed for BLS, named group sparse regularization for broad multitask learning system using related task-wise (BMtLS-RG) is proposed, significantly boosting BLS’s ability to generalize in MTL environments.
- 5
- 5
- Enhancing the Performance of Neural Networks Through Causal Discovery and Integration of Domain Knowledge5
A generic methodology to encode hierarchical causality structure among observed variables into a neural network in order to improve its predictive performance and an ablation study underscores the value of integrating structural and quantitative causal knowledge in enhancing the neural network's predictive performance incrementally.
- 5
- 5
- 5
- Bilinear Probabilistic Canonical Correlation Analysis via Hybrid Concatenations5
BPCCA is proposed, a new bilinear extension of PCCA, by introducing a hybrid joint model that preserves matrix structures indirectly via hybrid vector-based and matrix-based concatenations and enables BPCCA to gain more model flexibility in capturing two-view correlations and obtain close-form solutions in parameter estimation.
- 5
- 5
- 5
- 5
- 5
- A New Fragile Mesh Watermarking Algorithm for Authentication5
A new fragile watermarking algorithm based on the global characteristics of the mesh geometry to authenticate 3D mesh models is proposed, which shows that the embedded watermark is invariant to affine transformation but sensitive to other operations.
- 5
- 5
- A Divide-and-Conquer Learning Approach to Radial Basis Function Networks5
This paper presents a new divide-and-conquer based learning approach to radial basis function (RBF) networks, in which a conventional RBF network is divided into several RBF sub-networks, and finds that the performance generally varies with the different decompositions of the network’s input and the hidden layer.
- 5
- A Dual Structural Radial Basis Function Network for Recursive Function Estimation5
A new variant of extended normalized RBF (ENRBF) network is proposed to implement each subRBF net, which uses an adjustable p-order single-term polynomial, rather than the first-order one, to fit the relations between each hidden unit and each output unit.
- 5
- The MSE reconstruction criterion for independent component ordering in ICA timeseries analysis.5
This paper proposes to order the components according to their contribution in the mean square error between the observed time series and the reconstruction from these components, which leads to a typical combinational optimization problem.
- AnchorMoE: Interpretable Time Series Classification via Anchor-Routed MoE4
This paper proposes AnchorMoE, an interpretable-by-construction classification framework built upon a Mixture-of-Experts (MoE) architecture, which achieves highly competitive classification performance while faithfully grounding its decisions in the raw time series.
- 4
- Unknown-Aware Bilateral Dependency Optimization for Defending Against Model Inversion Attacks4
The bilateral dependency optimization strategy (BiDO) is proposed, a dual-objective approach that minimizes the dependency between input features and latent representations, while simultaneously maximizing the dependency between latent representations and labels.
- MOTEA-II: A Collaborative Multiobjective Transformation-Based Evolutionary Algorithm for Bilevel Optimization4
Experimental results show that the proposed collaborative MOTEA-II can achieve performance comparable to that of the previous MOTEA and three other representative EA-based bilevel optimization approaches, but using much fewer computational resources.
- 4
- Online Heterogeneous Feature Selection4
To provide a reliable foundation for evaluating feature subsets under dynamic and heterogeneous data streams, an incremental graph-unified metric (IGUM) is introduced that mitigates information loss between heterogeneous features by leveraging graph structures to unify feature-value-level and interfeature-level relationships.
- MOKD: Cross-domain Finetuning for Few-shot Classification via Maximizing Optimized Kernel Dependence4
A bi-level optimization framework to learn a set of class-specific representations that match the cluster structures indicated by labeled data of the given task, and can not only achieve better generalization performance on unseen domains in most cases but also learn better data representation clusters.
- Uniform Tensor Clustering by Jointly Exploring Sample Affinities of Various Orders4
The proposition of employing high-order affinities to characterize multiple sample relationships as a strategic means to circumnavigate the concentration effects is advanced and a novel clustering method named uniform tensor clustering (UTC) is proposed, which learns a consensus low-dimensional embedding for clustering by the synergistic exploitation of multiple-order affinities.
- 4
- A new grouping strategy-based hybrid algorithm for large scale global optimization problems4
A new variable grouping strategy which can be applicable to fully non-separable LSGO problems is proposed, and a new line search method is designed which can make a quick scan to arrive in promising regions and help the newVariable grouping strategy to divide the LSGO problem properly.
- 4
- 4
- Kernel Fuzzy Similarity Measure-Based Spectral Clustering for Image Segmentation4
A novel similarity measure called kernel fuzzy similarity measure is proposed first, Then this novel measure is integrated into spectral clustering to get a new clustering method: Kernel fuzzy similarity based spectral clustered (KFSC).
- 4
- [Identifying the characteristics of FTIR spectra of herba epimedii icariin via wavelet analysis and RBF neural network].4
Large variations in the spectrum caused by low content of icariin, weak absorption peaks and noise were successfully removed by applying correlation coefficient comparison and multilevel wavelet analysis, which significantly increased the quality of classification of RBF neural network.
- A THEORETICAL ANALYSIS OF MULTISCALE ENTROPY UNDER THE INVERSE GAUSSIAN DISTRIBUTION4
The theoretical basis of MSE is presented under the inverse Gaussian distribution, which is a typical model for physiological, physical and financial data sets, providing a reliable foundation for potential applications of M SE to explore complex physical and physiological time series.
- Writer Identification Using a Hybrid Method Combining Gabor Wavelet and Mesh Fractal Dimension4
A new hybrid method combining Gabor function and mesh fractal dimension for off-line, text-independent writer identification is proposed, capable of extracting more useful features to distinguish the handwritings.
- A Maximum Weighted Likelihood Approach to Simultaneous Model Selection and Feature Weighting in Gaussian Mixture4
This paper identifies the clustering structure and the relevant features automatically and simultaneously in the context of Gaussian mixture model by introducing two sets of weight functions under the recently proposed Maximum Weighted Likelihood (MWL) learning framework.
- 4
- 4
- Further studies on temporal factor analysis: comparison and Kalman filter-based algorithm4
This paper implements the TFA algorithm provided in Xu (1998, 2000), and compares it with non-temporal one to show the importance of consid ering temporal relationship in factor analysis, and sets up a connection between the T FA andtrad itional %ltering problems in control theory, and presents an alternative T FA algorithm.
- 4
- 4
- PI-H2T: Enhancing Long-Tailed Visual Recognition With Permutation-Invariant and Head-to-Tail Feature Fusion3
PI-H2T enhances the representation space through permutation-invariant representation fusion (PIF), yielding more clustered features and automatic class margins, and adjusts the biased classifier by transferring semantic information from head to tail classes via head-to-tail fusion (H2TF), improving tail class diversity.
- Epistemic Uncertainty for Generated Image Detection3
A novel framework for AI-generated image detection through epistemic uncertainty, aiming to address critical security concerns in the era of generative models by utilizing advanced large-scale vision models pre-trained on extensive natural images to estimate the epistemic uncertainty of images and flag those with high uncertainty as generated.
- 3
- Implementing Trust in Non-Small Cell Lung Cancer Diagnosis with a Conformalized Uncertainty-Aware AI Framework in Whole-Slide Images3
It is illustrated that an AI model wrapped with TRUECAM significantly outperforms models that lack such guidance, in terms of classification accuracy, robustness, interpretability, and data efficiency, while also achieving improvements in fairness.
- 3
- An Expensive Multi-Objective Optimization Algorithm Based on Decision Space Compression3
Experimental studies show that the proposed algorithm achieves a good performance in handling expensive multi-objective optimization problems with high-dimensional decision space.
- 3
- A Total-variation Constrained Permutation Model for Revealing Common Copy Number Patterns3
This paper proposes to tackle the issue of recovering the hidden blocks within single-cell DNA-sequencing data through continuous sample permutations such that similar samples are positioned adjacently, and demonstrates its ability to recover the hidden structures of single- cell DNA sequences.
- 3
- 3
- Modeling spatial relations of human body parts for indexing and retrieving close character interactions3
A new human pose representation, called Spatial Relations of Human Body Parts (SRBP), to represent spatial relations between body parts of the subject(s), which intuitively describes how much the body parts are interacting with each other.
- Expanding dictionary for robust face recognition: pixel is not necessary while sparsity is3
It is made to the conclusion that solving the approximation problem in raw pixel space is not necessary for robust face recognition, while solving in a subspace with a much smaller feature dimensionality is enough when the dictionary is well expanded.
- 3
- 3
- Active contours with a joint and region‐scalable distribution metric for interactive natural image segmentation3
This study presents an efficient active contour with a joint and region-scalable distribution metric for interactive natural image segmentation that can deal with the complex appearance and intensity in homogeneity and has the advantages of fast convergence and easy implementation.
- 3
- 3
- 3
- 3
- 3
- 3
- 3
- A Learning Framework for Blind Source Separation Using Generalized Eigenvalues3
This paper presents a learning framework for blind source separation (BSS), in which the BSS is formulated as generalized Eigenvalue (GE) problem, where the unknown unmixing matrix directly works out from the GE equation without timeconsuming iterative learning.
- 3
- A Fast Implementation of Radial Basis Function Networks with Application to Time Series Forecasting3
The DCRBF network is a hybrid system consisting of several sub-RBF networks, each of which takes a sub-input space as its input, which can considerably reduce the structural complexity of a RBF network, whereby the net’s learning becomes much faster.
- 3
- 3
- 3
- How to Ask the AI: A User Perspective Survey for Large Language Model Prompting2
This article provides actionable guidelines for formulating effective LLM prompts across diverse real-world tasks and contributes by developing an intuitive evaluation strategy for prompt efficacy; providing prompting workflow demonstrations on representative applications; and maintaining a dynamically updated open-source project to ensure the core takeaways remain up-to-date.
- 2
- Convolution Filter Compression via Sparse Linear Combinations of Quantized Basis2
A novel compression method that generates the convolution filters in each layer using a set of learnable low-dimensional quantized filter bases and reconstructs the convolution filters by stacking the linear combinations of these filter bases.
- Taking a Part for the Whole: An Archetype-agnostic Framework for Voice-Face Association2
An Archetype-agnostic Subspace Merging method is carefully designed to perform feature calibration which can well get rid of the archetype dependence to facilitate the mutual perception of datas and can well associate the voice-face data to benefit various kinds of cross-modal cognitive tasks.
- 2
- Spatial–temporal interaction module for action recognition2
An effective spatial–temporal interaction (STI) module is proposed that outperforms the existing counterparts on benchmark datasets and the computation cost of the proposed method is far less than 3D convolution.
- Cross-domain Prototype Learning from Contaminated Faces via Disentangling Latent Factors2
A disentangled heterogeneous prototype learning framework, dubbed DisHPL, which consists of one encoder-decoder generator and two discriminators, which aims to learn the variation-free target domain prototype for a contaminated input image from the source domain and meanwhile preserve the personal identity.
- Federated-PCA on Vertical-Partitioned Data2
A Federated-PCA machine learning approach, in which the PCA method is used to reduce the dimensionality of sample data for all parties and extract the principal component feature information to improve the efficiency of subsequent training work.
- 2
- Lip Password-Based Speaker Verification Without a Priori Knowledge of Speech Language2
This work proposes a novel visual speaker verification approach based on lip password without a priori knowledge of speech language, i.e. unknown language alphabet that takes advantage of the diagonal structure of sparse representation to preserve the temporal order of lip sequences by employ a diagonal-like mask in pooling stage and build a pyramid spatiotemporal features containing the structural characteristic under lip password.
- Optimization of Spectrum-Energy Efficiency in Heterogeneous Communication Network2
To improve the efficiency of optimization algorithm, an alternate optimization algorithm is presented to address the proposed complex MaOP and simulation results show the effectiveness of the algorithm.
- Texture Image Segmentation Using Spectral Clustering2
The Self-tuning Spectral Clustering (SSC) is used to the representative samples to do the clustering of texture image segmentation to improve the efficiency and avoid dependence on parameters.
- Video color conceptualization using optimization2
This paper presents a novel method to force neighboring pixels with similar intensities to have similar color, and formalizes the color conceptualization as an optimization problem with a quadratic cost function.
- Online Learning towards Big Data Analysis in Health Informatics2
This paper introduces online learning and proposes the method for data mining of big data in health informatics, where online learning will preform the data analysis dynamically by the time the data are generated.
- Feature Weighted Kernel Clustering with Application to Medical Data Analysis2
This paper studies the kernel-based clustering method which can conduct nonlinear partition on input patterns and addresses two challenging issues in unsupervised learning environment: feature relevance estimate and cluster number selection.
- Secure Watermarking on 3D Geometry via ICA and Orthogonal Transformation2
This paper applies the technique of independent component analysis to improve the security of 3D geometry watermarking by using the ICA de-mixing matrix for transformation with an orthogonal transformation matrix.
- An Adaptive Image Watermarking Scheme Using Non-separable Wavelets and Support Vector Regression2
The superiority of non-separable wavelet transform (DNWT) in capturing image features combined with the good generalization ability of support vector regression provide us with a promising way to design a more robust watermarking algorithm featuring a better trade-off between the robustness and imperceptivity, the main duality of watermarked algorithms.
- Information hiding for media authentication and covert communication2
A sequential quantization strategy (SQS) is proposed to improve the sensitivity of the embedded watermark to illegal modifications while preserving the property of semi-fragility, e.g. resistance to some content-preserving processing.
- 2
- 2
- A maximum likelihood approach to temporal factor analysis in state-space model2
This paper builds a connection between TFA and a traditional filtering problem in control theory, through which a new adaptive TFA algorithm is proposed and analyzed its solution properties accordingly.
- A Novel Markov Pairwise Protein Sequence Alignment Method for Sequence 665 Comparison2
A novel Markov pairwise protein sequence alignment (MPPSA) method that takes the local context dependencies into consideration and the numerical results have shown its superiority to the Smith-Waterman algorithm for pairwiseprotein sequence comparison.
- 2
- 2
- A New Information-Theoretic Based Ica Algorithm For Blind Signal Separation2
A flexible nonlinearity function that is a single polynomial term with the exponent learnable is considered, the separation ability of this function is analysed, and a new ICA algorithm is proposed that can successfully separate the mixture of sub- Gaussian and super-Gaussian sources.
- A DIVIDE-AND-CONQUER FAST IMPLEMENTATION OF RADIAL BASIS FUNCTION NETWORKS WITH APPLICATION TO TIME SERIES FORECASTING2
This paper presents a new divide-and-conquer learning approach to radial basis function networks (DCRBF) and has experimentally shown its outstanding learning performance on forecasting two real time series as well as synthetic data in comparison with a conventional RBF one.
- 2
- HyReaL: Clustering Attributed Graph via Hyper-complex Space Representation Learning1
It turns out that the node representations learned by HyReaL are more discriminative and widely suit downstream clustering with different $k$s, naturally alleviating the OS problem.
- Heterogeneous neural blind deconvolution: A signal processing-empowered foundation feature extractor for bearing fault diagnosis1
A theoretically well-grounded approach to build an enhanced backbone model for various key downstream tasks, and proves that incorporating neural heterogeneity by combining linear and quadratic neurons leads to a more efficient approximation of any univariate real coefficient polynomial.
- 1
- Adaptive Point-Prompt Tuning: Fine-Tuning Heterogeneous Foundation Models for 3D Point Cloud Analysis1
This paper proposes the Adaptive Point-Prompt Tuning (APPT) method, which fine-tunes pre-trained models with a modest number of parameters, enabling direct point cloud processing without heterogeneous mappings.
- Evo-TFS: Evolutionary Time-Frequency Domain-Based Synthetic Minority Oversampling Approach to Imbalanced Time Series Classification1
Evo-TFS is proposed, a novel evolutionary oversampling method that integrates both time- and frequency-domain characteristics and outperforms existing oversampling methods, significantly enhancing the performance of time-domain and frequency-domain classifiers.
- Modeling and Identifying Distractors with Curriculum for Robust 3D Gaussian Splatting1
This work proposes to learn a soft mask of the likelihood of the distractors, and develops techniques to model the spatial pattern of distractors and learn them from a design curriculum, avoiding confusion between clean and noisy pixels.
- AI in the era of large language models: current state, applications, limitations, and future1
How AI has entered a new era of "Daily Conversation", guides you through its powerful applications and risks, and finally looks ahead to a future of AI.
- 1
- 1
- 1
- 1
- 1
- 1
- 1
- 1
- 1
- On incremental collaborative appearance model and regional particle filtering for lip region tracking1
This paper presents an efficient lip region tracking approach in virtue of incremental collaborative appearance model and regional particle filtering, and proposes a regional particle filtering for motion state estimation, by taking scale size and rotation estimation, reducing the computational load and alleviating the tracking drift.
- Robust Heterogeneous Discriminative Analysis for Single Sample Per Person Face Recognition1
This work proposes a new patch-based method, namely Robust Heterogeneous Discriminative Analysis (RHDA), to tackle single sample face recognition with SSPP, and develops a novel Fisher-like criterion, which incorporates two manifold embeddings, to learn heterogeneous discriminative representations of image patches.
- An Evolutionary Algorithm for TD-LTE Resource Allocation Based on Adaptive Fairness Threshold1
A TD-LTE resource allocation evolutionary algorithm based on adaptive fairness threshold is proposed to solve this model through introducing adaptive fairness thresholds and empirical studies show that the proposed model and algorithm can effectively decrease the cross time slot interference and improve the transmit rate of OFDM system under the same fairness.
- Outlier Detection Based on Local Kernel Regression for Instance Selection1
An outlier detection approach based on local kernel regression for instance selection that evaluates the reconstruction error of instances by their neighbors to identify the outliers is proposed.
- 1
- 1
- 1
- A Novel Automatic Lip Reading Method Based on Polynomial Fitting1
This paper addresses the problem of isolate number recognition using visual information only by utilizing the intensity transformation and spatial filter to estimate the minimum enclosing rectangle of mouth in each frame and presenting a method to recognize the speech based on the polynomial fitting.
- 1
- Topological Self-Similar Networks Introduced by Diffusion-Limited Aggregation Mechanism1
This model introduces multiplicative growth, aging and geographical preferential attachment mechanisms, whereby featuring topological self-similar property and hierarchical modularity in fractal geometries in the viewpoint of network.
- 1
- A Batch Rival Penalized EM Algorithm for Gaussian Mixture Clustering with Automatic Model Selection1
This paper develops a batch RPEM algorithm that need not assign the learning rate analogous to the EM, but still preserve the capability of automatic model selection, and the convergence speed of thisbatch RPEM is faster than the EM and the adaptive RPEM.
- 1
- 1
- 1
- Discrimination of Patchoulis of Different Geographical Origins with Two-Dimensional IR Correlation Spectroscopy and Wavelet Transform1
The results in this paper showed that it’s quite easy to use two-dimensional infrared (2D IR) correlation spectra, especially after applying wavelet transform process, to discriminate patchouli, and Fourier transform infrared spectroscopy would become more powerful in analysis and discrimination.
- 1
- 1
- 1
- 1
- –
- Beyond Statistical Co-occurrence: Unlocking Intrinsic Semantics for Tabular Data Clustering–
This paper proposes Tabular-Augmented Contrastive Clustering (TagCC), a novel framework that anchors statistical tabular representations to open-world textual concepts via semantic-aware transformation and enriches the statistical tabular representations with the open-world semantics encapsulated in these anchors.
- –
- –
- –
- –
- –
- Correlative Fusion-Based Multi-Instance Partial Multi-Label Learning–
Correlation-Fusion MIPML (CF-MIPML), an innovative framework that integrates Label Confidence Generation (LCG) and Candidate Label Disambiguation (CLD), outperforms existing state-of-the-art methods, enhancing handling of complex label ambiguity and improving overall model generalization in practical MIPML scenarios.
- –
- –
- –
- DRFGD: Disentangled Representation-Focused Generative Defense for Attack-Tolerant Cross-Modal Hashing–
A novel Disentangled Representation-Focused Generative Defense framework for attack-tolerant cross-modal hashing that substantially improves retrieval robustness under various attacking scenarios, and shows its improved defense performance in comparison with the SOTA works.
- –
- –
- –
- Semantic-Guided Fine-Tuning of Foundation Model for Long-Tailed Visual Recognition–
A novel distribution mismatch-aware compensation factor is proposed, which is specifically designed to rectify the prediction bias caused by the ignored inconsistent distribution based on the authors' theoretical analysis, and is seamlessly integrated into the loss function.
- –
- –
- –
- –
- –
- –
- –
- –
- Exploiting the Complementarity of Bilateral Domains for Fast Lane Detection–
A novel network that smoothly introduces frequency space into the spatial domain to achieve accurate and fast lane detection and significantly outperforms the state-of-the-art approaches is proposed.
- –
- –
- –
- –
- –
- –
- –
- Federated Principal Component Analysis for Vertically Partitioned Data–
VFedPCA and VFedAKPCA are introduced, pioneering federated algorithms designed for linear and nonlinear PCA in such distributed environments, to facilitate collaborative PCA computations across distributed clients while preserving data privacy by avoiding raw data exchanges.
- EvoSampling: A Granular Ball-based Evolutionary Hybrid Sampling with Knowledge Transfer for Imbalanced Learning–
This paper proposes an evolutionary multi-granularity hybrid sampling method, called EvoSampling, and develops a granular ball-based undersampling method that removes noise in a multi-granular fashion, thereby enhancing data quality.
- –
- –
- –
- Compact Neural Networks via Stacking Designed Basic Units–
A new method termed TissueNet is presented, which directly constructs compact neural networks with fewer weight parameters by independently stacking designed basic units, without requiring additional judgement criteria anymore, which provides a new promising way for network compression.
- –
- –
- A Scoring Model Assisted by Frequency for Multi-Document Summarization–
A scoring model which combines position information with frequency to identify the salience of sentences is proposed, which outperforms the other state-of-the-art extractive models.
- How to Make a Query in Image Retrieval with Partial Information Extracted from Multiple Image Samples?–
A novel framework is presented that allows a user to make an image query composed of several partial information extracted from multiple image samples via Boolean operations (i.e. AND, OR and NOT).
- –
- –
- –
- –
- –
- –
- A New Adaptive Hybrid Algorithm for Large-Scale Global Optimization–
A new adaptive hybrid algorithm (NAHA) for LSGO problems, which integrates the global search, local search and grouping search and greatly improves the search efficiency and designs an automatic resource allocation strategy which can allocate resources to different optimization strategies automatically and adaptively according to their performance and different stages.
- –
- Computational Intelligence in Astronomy–
This special issue is mainly dedicated to the workshop of AI in astronomy, in conjunction with the 12th International Conference on Computational Intelligence and Security held in Wuxi, China on December 16-19, 2016, which has 136 final accepted papers selected from 288 online submissions.
- –
- –
- –
- Anonymized Distance Filter in Hamming Space–
An anonymized distance filter that can test and rank instances in a Hamming-ball search without knowing explicit distance values is designed and demonstrated on both simulated and real data sets.
- –
- Nature-Inspired Algorithms and Applications: Selected Papers from CIS2013–
This material is protected by copyright and other intellectual property rights, and duplication or sale of all or part of any of the repository collections is not permitted, except that material may be duplicated by you for your research use or educational purposes in electronic or print form.
- –
- Proximal Average Approximated Incremental Gradient Method for Composite Penalty Regularized Empirical Risk Minimization–
A new PA-based algorithm called PA-SAGA is proposed, which is optimal in both convergence rate and per-iteration cost, by incorporating into incremental gradient-based framework.
- –
- –
- –
- –
- –
- –
- –
- A Cooperative and Penalized Competitive Learning Approach to Gaussian Mixture Clustering–
This paper proposes a novel learning algorithm called Cooperative and Penalized Competitive Learning (CPCL), which implements the cooperation and penalization mechanisms simultaneously in a single competitive learning process to have good convergence speed, precision and robustness.
- On feature selection, kernel learning and pairwise constraints for clustering analysis–
A new training paradigm in which local models are coupled by sharing a nonnegative feature weight vector among them is proposed, which is theoretically guaranteed not only be able to produce concise local model in each neighborhood, but also promote consistency among the local models.
- –
- –
- A New Quantifying Crossover With Ternary Representation–
A new ternary representation is proposed through which the survival and construction of a schema for any crossover operator can be distinguished easily, and the probabilities that a schema will survive and be constructed from other schemata can be estimated for an arbitrary crossover.
- Hong Kong Chapter Report [Family Corner]–
The IEEE (Hong Kong) Computational Intelligence Chapter was founded in 2003 and aims to coordinate and support work and to arouse the public’s interests in the field of Computational intelligence in the city as well as the region.
- –
- Classifying G-Protein Coupled Receptors with Hydropathy Blocks and Support Vector Machines–
A new method based on features generated from the hydropathy properties of the amino acid sequences to recognize G-Protein Coupled Receptors (GPCRs) belonging to the same families or subfamilies is developed.
- –
- –
- –
- –
- Expectation-MiniMax Approach to Clustering Analysis–
A general approach named Expectation-MiniMax (EMM) for clustering analysis without knowing the cluster number that provides a better alternative way to optimize the EM contrast function with at least two advantages: faster model parameter learning speed, and automatic model-complexity selection capability.
- –
- –
- –
- –
Publication data from OpenAlex, with missing venues and authors filled in from Crossref; citation counts are the higher of OpenAlex and Semantic Scholar, last synced 2026-10-15. One-sentence summaries under some papers are written by Semantic Scholar’s model. Citation counts may be lower than on Google Scholar, which indexes more sources.
Report an error
Wrong papers, two people merged into one, or a profile that should not be here? Tell us and we will fix or hide it.