Is this you? Claim this profile to correct it, add a bio and choose the work people see first.
Claim this profileWorks224 from public data
- 1,273
- Pythagorean fuzzy MCDM method based on CoCoSo and CRITIC with score function for 5G industry evaluation301
The main advantages of proposed algorithm are: (1) have no counterintuitive phenomena; (2) without division or antilogarithm by zero problem; (3) own stronger ability to distinguish alternatives.
- 187
- 147
- 113
- 113
- Learning to Purification for Unsupervised Person Re-Identification100
This work proposes to take advantage of the knowledge of teacher model in an offline scheme to purify the label noise, and achieves a state-of-the-art accuracy 85.8% @mAP and 94.5% @Rank-1 on the challenging Market-1501 benchmark with ResNet-50 under the fully unsupervised setting.
- A Novel Learnable Dictionary Encoding Layer for End-to-End Language Identification79
A novel learnable dictionary encoding layer is proposed in this paper that can accumulate high-order statistics from variable-length input sequence and generate an utterance level fixed-dimensional vector representation.
- 73
- Dependency-based long short term memory network for drug-drug interaction extraction71
A dependency-based deep neural network model that can learn effective relation information with less feature engineering in the task of DDI extraction is proposed, and the experimental results show that the model excels at balancing the Precision and Recall values.
- Dual Transfer Learning70
This paper proposes a novel approach, Dual Transfer Learning (DTL), which simultaneously learns the marginal and conditional distributions, and exploits the duality between them in a principled way.
- Automated evaluation of building alignments in generalized maps51
This article firstly analyses the requirements (constraints) related to the generalization of building alignments, then, it focuses on three more specific constraints, i.e. on existence, orientation of alignments and spatial distribution of composing buildings.
- 49
- Maximum Mean and Covariance Discrepancy for Unsupervised Domain Adaptation49
An unsupervised learning model based on MMCD abbreviated as McDA was proposed and efficiently optimized to resolve the domain adaptation problem, and experiments on image classification show that McDA outperforms other representative domain adaptation methods, which implies the effectiveness of MMCD in domain adaptation.
- 45
- Gene Ranking of RNA-Seq Data via Discriminant Non-Negative Matrix Factorization43
DNMF is suggested to be an effective method for the analysis of differential gene expression and gene ranking for RNA-seq data, substantially outperforming all other benchmarked methods.
- 41
- CRD: Fast Co-clustering on Large Datasets Utilizing Sampling-Based Matrix Decomposition.40
A general framework for fast co-clustering large datasets, CRD is proposed, which achieves competitive accuracy but with much less computational cost by utilizing recently developed sampling-based matrix decomposition methods.
- CRD40
A general framework for fast co-clustering large datasets, CRD is proposed, which achieves competitive accuracy but with much less computational cost by utilizing recently developed sampling-based matrix decomposition methods.
- Characterization and detection of building patterns in cartographic data: Two algorithms38
This paper studies existing topographic maps at large to medium scales, and proposes and discusses a comprehensive typology of building patterns, their distinctions and characteristics, which includes linear alignments, curvilinear, align-along-road alignments and nonlinear clusters.
- 37
- Supervised Mineral Prospectivity Mapping via Class-Balanced Focal Loss Function on Imbalanced Geoscience Datasets33
A novel class-balanced focal loss function (CBFL) is provided to address the training data imbalance issue in deep learning modeling for MPM and it is demonstrated that the CBFL outperformed the other loss functions.
- Universum Prescription: Regularization Using Unlabeled Data33
It is shown that simply prescribing "none of the above" labels to unlabeled data has a beneficial regularization effect to supervised learning.
- An Integrated Framework for Data-Driven Mineral Prospectivity Mapping Using Bagging-Based Positive-Unlabeled Learning and Bayesian Cost-Sensitive Logistic Regression32
An integrated framework for data-driven MPM is proposed, which is denoted as BPUL–BCSLR, with Bagging-based positive-unlabeled learning (BPUL) and Bayesian cost-sensitive logistic regression ( BCSLR) in order to address simultaneously the issues mentioned above.
- 32
- InterBN: Channel Fusion for Adversarial Unsupervised Domain Adaptation32
A novel unsupervised domain adaptation (UDA) method based on Interchangeable Batch Normalization (InterBN) to fuse different channels in deep neural networks for adversarial domain adaptation to drastically increase the domain adaption ability.
- A generalist foundation model and database for open-world medical image segmentation31
A publicly accessible database is constructed, MedSegDB, based on a tree-structured hierarchy and annotated from 129 public medical segmentation repositories and 5 in-house datasets, indicating that MedSegX achieves state-of-the-art performance across various modalities and organ systems in in-distribution (ID) settings.
- 30
- Deep Co-Clustering29
A deep learning model for co-clustering called DeepCC, which utilizes the deep autoencoder for dimension reduction, and employs a variant of Gaussian Mixture Model (GMM) to infer the cluster assignments and a mutual information loss is proposed to bridge the training of instances and features.
- Informative pairs mining based adaptive metric learning for adversarial domain adaptation28
This paper proposes an informative pairs mining based adaptive metric learning (IPM-AML), where a novel two-triplet-sampling strategy is advanced to select informative positive pairs from the same classes and informative negative pairs from different classes, and a metric loss imposed with special weights is further utilized to adaptively pay more attention to those more informative pairs which can adaptively improve discrimination.
- Enhancing Stock Price Trend Prediction via a Time-Sensitive Data Augmentation Method27
A simple yet effective time-sensitive data augmentation method for stock trend prediction by corrupting high-frequency patterns of original stock price data as well as preserving low-frequency ones in the frame of wavelet transformation.
- Semi-Supervised Projective Non-Negative Matrix Factorization for Cancer Classification26
A semi-supervised projective non-negative matrix factorization method to learn an effective classifier from both labeled and unlabeled samples, thus boosting subsequent cancer classification performance and developed a multiplicative update rule (MUR) to optimize Semi-PNMF and proved its convergence.
- Prefix Conditioning Unifies Language and Label Supervision25
Novel Prefix Conditioning is introduced, a simple yet effective method that helps disentangle dataset biases from visual concepts and improves zero-shot image recognition and robustness to image-level distribution shift.
- 25
- 24
- 23
- Enhancing the association in multi‐object tracking via neighbor graph23
This study proposes to make full use of the neighboring information via GCN in MOT for the first time, and can remarkably reduce the identity switches whilst achieve state‐of‐the‐art overall performance.
- Discriminative Locality Preserving Canonical Correlation Analysis23
A novel dimension reduction method based on canonical correlation analysis, called discriminative locality preserving canonical correlations analysis (DLPCCA) is proposed, which outperforms the representative algorithms including CCA, KCCA, LPCCA and LDCCA.
- 22
- 22
- FIXED: Frustratingly Easy Domain Generalization with Mixup21
This work proposes a simple yet effective enhancement for Mixup-based DG, namely domain-invariant Feature mIXup (FIX), which learns domain- Invariant representations for Mix up and significantly outperforms nine state-of-the-art related methods.
- Another Robust NMF: Rethinking the Hyperbolic Tangent Function and Locality Constraint21
A robust NMF termed as tanhNMF for short is proposed, which rethinks the hyperbolic tangent (tanh) function as a robust loss to evaluate the reconstruction error and devise a locality constraint to regularize tanh NMF to model data locality.
- 21
- Discriminative Geometric-Structure-Based Deep Hashing for Large-Scale Image Retrieval20
This article proposes a discriminative geometric-structure-based deep hashing method (DGDH), which investigates three novel loss terms based on class centers to induce the so-called discrim inative geometrical structure, and theoretically analyze the robustness and generalization of the proposed method.
- 19
- 19
- Representing Vector Geographic Information As a Tensor for Deep Learning Based Map Generalisation18
This article extracts some representation issues from a literature review and proposes different ways to represent vector geographic information as a tensor, and demonstrates the interest of some of the propositions with experiments that show a visual improvement for the generation of generalised topographic maps in urban areas.
- Person re-identification via adaptive verification loss18
This paper develops an adaptive verification loss, termed as ADV-Loss, which empowers the popular verification loss with the flexibility of adaptively addressed the hard triple units and learns a re-weighted strategy from the triplet loss to balance the sample pairs.
- The evalutation of spatial distribution density in map generalization18
An object-oriented density measurement based on skeletonization of gap space is proposed so that density can be calculated with each object and its spatial territory, in contrast to other approaches.
- IoUformer: Pseudo-IoU prediction with transformer for visual tracking17
A novel simple tracking paradigm based on the intersection over union (IoU) value prediction, which can predict reliable IoU values near-linear with the ground truth, which paves a safe way for the new IoU-based siamese tracking paradigm.
- Cascade Semantic Fusion for Image Captioning17
A cascade semantic fusion architecture to mine the representative features to encode image content through attention mechanism without bells and whistles and shows that the CSF can assist image captioning model in selecting the object regions of interest.
- WiHumidity: A Novel CSI-Based Humidity Measurement System17
A novel system, namely WiHumidity, is provided, which utilizes the special diversity of channel state information (CSI) to alleviate multipath effect at the receiver and can achieve 79% measurement accuracy.
- 15
- A generic MOT boosting framework by combining cues from SOT, tracklet and re-identification15
A generic boosting framework for multiple object tracking (MOT) that combines the techniques of single object tracking, tracking fragment (tracklets) and re-identification mechanism through casting them into a multi-label energy optimization and then innovatively solving it using the α-\documentclass[12pt]{minimal}.
- Analysis of Spatial Data15
The development of simultaneous spatial-temporal autoregressive model and validate the asymptotic property of maximum likelihood estimates under such model are developed and the establishment of a new nonparametric Bayesian prior and corresponding computational algorithm is established.
- 14
- Object-aware semantics of attention for image captioning14
The object- Aware semantic attention object-aware semantic attention (OSA) based captioning model allows the explicit associations between the objects by coupling the attention mechanism with three types of semantic concepts, i.e., the category information, relative sizes of the objects, and relative distances between objects.
- HICC: an entropy splitting-based framework for hierarchical co-clustering14
A new co-clustering algorithm, HICC, with solid theoretical background is presented, which simultaneously constructs a hierarchical structure of both row and column clusters, which retains sufficient mutual information between rows and columns of the contingency table.
- 14
- LanP: Rethinking the Impact of Language Priors in Large Vision-Language Models13
A benchmark called LanP is proposed to rethink the impact of Language Priors in LVLMs, designed to investigate how strong language priors are in current LVLMs.
- Semantics-Enriched Cross-Modal Alignment for Complex-Query Video Moment Retrieval13
This paper presents a novel solution termed semantics-enriched video moment retrieval method (SVMR), which can effectively and explicitly model the hierarchical multi-granularity semantics of complex textual query and explores cross-token relations to offer multiple granularity query representations with hierarchical semantic contexts of semantically associated tokens.
- 13
- Joint high-resolution feature learning and vessel-shape aware convolutions for efficient vessel segmentation12
A new topology- and shape-aware model named Multi-branch Vessel-shaped Convolution Network (MVCN) to adaptively learn high-resolution representations from retinal vessel imagery and thereby capture high-quality topology and shape information thereon is proposed.
- Enhancing Online UAV Multi-Object Tracking with Temporal Context and Spatial Topological Relationships12
A temporal feature aggregation module (TFAM) is proposed, which effectively combines temporal contexts to obtain rich feature response maps in dynamic motion scenes to enhance the detection capability of the proposed tracker and a topology-integrated embedding module (TIEM) is introduced that captures the topological relationships between objects and their surrounding environment globally and sparsely, thereby integrating spatial layout information.
- 12
- 12
- Automated evaluation of generalized topographic maps12
This thesis has identified the following issues from the state-of-the-art of automated evaluation of generalization output: • Lack of formalized map requirements: cartographic constraints were specified at a knowledge level that is only human readable, thus not suitable for automated evaluation.
- Temporal-Enhanced Multimodal Transformer for Referring Multi-Object Tracking and Segmentation11
This study introduces a compact Transformer-based method, termed TenRMOT, that conducts feature fusion at both encoding and decoding stages to fully exploit the advantages of Transformer architecture and introduces a novel task called Referring Multi-Object Tracking and Segmentation and construct a new dataset named Ref-KITTI Segmentation.
- Phrase Grounding-Based Style Transfer for Single-Domain Generalized Object Detection11
This paper proposes a novel phrase-grounding-based style transfer (PGST) approach for single-domain generalized object detection that significantly improves mean average precision (mAP) and outperforms or matches the performance of domain-adaptive object detection methods, which require target domain data for training, in several challenging scenarios.
- GC-Flow: A Graph-Based Flow Network for Effective Clustering11
The resulting neural network, GC-Flow, retains the graph convolution operations while being equipped with a Gaussian mixture representation space, and enjoys two benefits: it not only maintains the predictive power of GCN, but also produces well-separated clusters, due to the structuring of the representation space.
- 11
- GateCap: Gated spatial and semantic attention model for image captioning11
A gated spatial and semantic attention captioning model (GateCap) which adaptively fuses spatial attention features with semantic attention features to achieve this goal and could reduce the side effect of the word-to-region misalignment at a time step over subsequent word prediction, thereby possibly alleviating emergence of incorrect words during testing.
- Enhancing unsupervised domain adaptation by discriminative relevance regularization11
A discriminative relevance regularization term (DRR) is proposed to enhance the performance of UDA by reducing the domain shift from the aspect of semantic relevance across domains by solving the min–max rank problem.
- Nonnegative Constrained Graph Based Canonical Correlation Analysis for Multi-view Feature Learning11
An improved CCA variant of Nonnegative Constrained Graph regularized CCA (NCGCCA) is proposed, which develops a nonnegative constrained graph based self-representation and offers a graph embedding schema to incorporate the underlying structure into CCA.
- Synthetic Over-sampling for Imbalanced Node Classification with Graph Neural Networks10
This work seeks to address the graph imbalance problem in graph neural networks by generating pseudo instances of minority classes to balance the training data, extending previous over-sampling-based techniques.
- Entity Profiling in Knowledge Graphs10
The distinctiveness of features is carefully measured by a HAS model, which is a scalable representation learning model to produce a multi-pattern entity embedding and shows that the approach facilitates human understanding of entities in KGs.
- 10
- 10
- Attention Focused Spatial Pyramid Pooling for Boxless Action Recognition in Still Images10
Experiments on two public action recognition benchmark datasets demonstrate that AttSPP-net can achieve promising results and even outweighs some methods based on ground-truth bounding boxes, and provides an alternative way towards practical applications.
- 10
- 9
- Self-Supervised Scene Dynamic Recovery from Rolling Shutter Images and Events9
An event-based SDR network within a self-supervised learning paradigm and an Event-based Inter/intra-frame Compensator (E-IC) is proposed to predict the per-pixel dynamic between arbitrary time intervals, including the temporal transition and spatial translation.
- 9
- 9
- 9
- 9
- Novel Approaches to Biomolecular Sequence Indexing.9
This survey describes some of the recent developments in biomolecular sequence indexing methods that allow efficient similarity search and focuses here on global similarity measures that compare sequences in full; such measures are important for comparing protein sequences and smaller biomolecules.
- 8
- 8
- 8
- Learning sequence-to-sequence affinity metric for near-online multi-object tracking8
A sequence-to-sequence affinity metric for the data association of near-online multi-object tracking that enhances the conventional track sequence to detection affinity metric and improves its robustness to object occlusion and missing.
- Online Multi-Modal Robust Non-Negative Dictionary Learning for Visual Tracking8
An online multi-modal robust non-negative dictionary learning (OMRNDL) algorithm to overcome this deficiency in conventional dictionary learning algorithms and validate the effectiveness of OMRNDL for visual tracking in both quantity and quality.
- 7
- Enabling Efficient and Privacy-Preserving Task Allocation with Temporal Access Control for Mobile Crowdsensing7
An Efficient and Privacy-Preserving Task Allocation with Temporal Access Control (EPTA-T) scheme for mobile crowdsensing that supports fine-grained and temporal access control through the utilization of an attribute-based access tree and function integration is proposed.
- Time-Varying Graph Learning Under Structured Temporal Priors7
This paper endeavors to learn time-varying graphs by using structured temporal priors that assume underlying relations between arbitrary two graphs in the graph sequence by proposing a structure named temporal graph to characterize the underlying real temporal relations.
- An Improved Differential Fault Analysis on Block Cipher KLEIN-647
An improved DFA is performed to KLEIN-64 and it is found that the differential propagation path and the distribution of the S -box can be fully utilized to distinguish the correct and wrong keys when a half-byte fault is injected in the 10 th round.
- Robust Normalized Squares Maximization for Unsupervised Domain Adaptation7
A novel normalized squares maximization (NSM) loss in which the maximum squares is normalized by the sum of squares of class sizes to explicitly increase the diversity of UDA predictions.
- 6
- SiamDF: Tracking training data-free siamese tracker6
A data-free siamese tracking algorithm named SiamDF is presented, which requires only a pre-trained backbone and no further fine-tuning on additional training data, and achieves impressive performance compared to well-established unsupervised and supervised tracking methods.
- 6
- 6
- Joint Modality Synergy and Spatio-temporal Cue Purification for Moment Localization6
A conceptually simple modality synergy strategy is explored to keep features modality specific while absorbing the other modality complementary information with both carefully designed cross-attention unit and non-contrastive learning.
- AttResNet6
This paper mainly refines the encoder with an attention-based ResNet model to provide better semantic features for the decoder and design an attention connection and then seamlessly couple it with the well-known ResNet.
- 6
- Inverse reinforcement learning for discrete-time linear systems based on inverse optimal control5
This paper develops a partially model-free IRL framework for reconstructing the cost function by introducing auxiliary control inputs and decomposing the algorithm into outer and inner loop, and develops a model-based IRL algorithm for reconstructing the cost function.
- 5
- 5
- 5
- 5
- 5
- Research on Sentiment Analysis Algorithm for Comments on Online Ideological and Political Courses5
Two deep learning algorithms were combined, i.e., a convolutional neural network (CNN) algorithm and a long short-term memory (LSTM) algorithm, to identify and analyze the emotional tendency of comments on online ideological and political courses to show that the recognition performance of the CNN+L STM-based sentiment analysis algorithm was better than that adopting the one-hot text vectorization method.
- 5
- Incremental Spectral Clustering via Fastfood Features and Its Application to Stream Image Segmentation5
An incremental spectral clustering method for stream data clustering and apply it to stream image segmentation that is able to learn data partitions incrementally and improve eigenvector approximation with more and more data seen from a stream.
- Two-Dimensional Euler PCA for Face Recognition5
Experimental results on popular face datasets show that 2D-ePCA outperforms the representative algorithms in terms of accuracy, computational overhead, and robustness.
- 5
- TransConv: Transformer Meets Contextual Convolution for Unsupervised Domain Adaptation4
Different from previous transformer based UDA architectures, TransConv has two special aspects: reviving the multilayer perception (MLP) of transformer encoders with Gaussian channel attention fusion for robustness, and mixing contextual features to highly efficient dynamic convolutions for cross-domain interaction.
- Takin-VC: Expressive Zero-Shot Voice Conversion via Adaptive Hybrid Content Encoding and Enhanced Timbre Modeling4
Experimental results demonstrate that the proposed Takin-VC method surpasses state-of-the-art zero-shot VC systems, delivering superior performance in terms of both speech naturalness and speaker similarity.
- 4
- Multi-Task Recurrent Modular Networks4
This work proposes multi-task recurrent modular networks (MT-RMN), which is a shared encoder and multiple task-specific decoders that recurrently operates over time, and aims to discover a set of generalizable sub-networks in the encoder that are assembled in different ways for different tasks.
- Correlation Self-Expression Shrunk for Subspace Clustering4
A correlation self-expression shrunk (CSS) is proposed to refresh LRR with two novel insights: the compound schatten -norm instead of the nuclear norm which is used to induce adaptive affinity matrix by addressing the issues of both sparse and dense representation and to model compact group structure and relieve side effect of the outliers.
- Transductive Nonnegative Matrix Tri-Factorization4
This paper proposes a transductive nonnegative matrix tri-factorization method (T- NMTF) to simultaneously exploit the label information of training examples and the statistical structure of features in the test domain, and proposes a manifold regularized T-NMTF (MT-NM TF) algorithm that exploits the local geometry structure of the dataset to boost discriminant power.
- 4
- Hierarchical co-clustering based on entropy splitting4
A new co-clustering algorithm with solid information theoretic background is presented which simultaneously constructs a hierarchical structure of both row and column clusters which retains sufficient mutual information between rows and columns of the contingency table.
- 4
- Application of Clustering Methods in Multivariate Data-Based Prospecting Prediction3
This study demonstrates the effectiveness of combining Ward clustering for variable transformation with TSC for the integrated analysis of categorical and numerical data, confirming its value in multi-source data research and its potential for further application.
- Event-Oriented State Alignment Network for Weakly Supervised Temporal Language Grounding3
The proposed Event-oriented State Alignment Network (ESAN) employs relative entropy for cross-modal alignment through knowledge distillation from pre-trained large models, thereby enhancing semantic coherence within each modality and ensuring consistency across modalities.
- 3
- 3
- 3
- 3
- Analyzing agricultural ecological efficiency in Weihai City based on the SBM, STIRPAT and SLM models3
- 3
- A Pruning Method Based on Feature Abstraction Capability of Filters3
A new pruning strategy, which aims to accelerate and compress off-the-shelf CNN models with weak Feature Abstraction Capability by combining the kernel sparsity of the filter with the dispersion of the feature maps activated by the filter.
- CrescendoNet: A New Deep Convolutional Neural Network with Ensemble Behavior3
CrescendoNet provides a new way to construct high performance deep convolutional neural networks with simple network architecture and allows a new path-wise training procedure, which can reduce the memory needed for training.
- 3
- 3
- Visualizing RDF Data Profile with UML Diagram3
A tool that can construct the actual schema, gather corresponding statistics, and can present a visual, UML diagram for RDF data sources, such as SPARQL endpoints or RDF dumps for datasets with complex structure is described.
- 3
- Probability output modeling for support vector machines3
This approach modeling the posterior probability output of multi-class SVMs is achieved by directly solving the equations that are based on the combination of the probability outputs of binary classifiers using the Bayes's rule.
- Task Allocation Method for Emergency Active Debris Removal Based on the Fast Elitist Non-Dominated Sorting Genetic Algorithm2
A Multi-type Chromosome Fast Elitist Non-Dominated Sorting Genetic Algorithm (MC-NSGA-II) is proposed to enable the satellite swarm to execute multiple coupled tasks in succession with improved optimization efficiency.
- 2
- 2
- Recovering Continuous Scene Dynamics from A Single Blurry Image with Events2
An Implicit Video Function is learned to represent a single motion blurred image with concurrent events, enabling the latent sharp image restoration of arbitrary timestamps in the range of imaging exposures, and a dual attention transformer is proposed to efficiently leverage merits from both modalities.
- 2
- 2
- Diff-Transfer: Model-based Robotic Manipulation Skill Transfer via Differentiable Physics Simulation2
This work introduces $\textit{Diff-Transfer}$, a novel framework leveraging differentiable physics simulation to efficiently transfer robotic skills, and proposes a novel path-planning method to generate sub-tasks, leveraging $Q-learning with a task-level state and reward.
- 2
- 2
- Visual analytics of genealogy with attribute-enhanced topological clustering2
This paper proposes an attribute-enhanced topological clustering method for exploring genealogy datasets based on partial least squares (PLS), and implements a visual analysis system with multi-view collaboration, including glyph, family tree view and parallel coordinate view, to represent, evaluate and explore the clustering features.
- Robust Similarity-Based Concept Factorization for Data Representation2
A similarity-based concept factorization (SCF) is proposed as a synthesis of the two typically important representative models, which enjoys similarity preservation, robustness to similarity perturbation, and ability of reconstructing samples.
- Neighbors-Based Graph Construction for Dimensionality Reduction2
This paper proposes a novel neighbors-based distance that measures the distance of two examples through their neighbors through theirNeighborhoods, and creates a virtual bridge point from the neighbors of each example and use it to link with others.
- Cross-Layer Convolutional Siamese Network for Visual Tracking2
A cross-layer convolutional Siamese network tracker which attempts to explore more semantic features of different layers from two aspects, and demonstrates the effectiveness of Siam-CC as compared to the state-of-the-art trackers.
- Boxless Action Recognition in Still Images via Recurrent Visual Attention2
A ReAttSPP-net is proposed for boxless action recognition that utilizes a recurrent neural network model of visual attention in order to extract information from a sequence of fixations and can achieve promising results and obtains higher recognition performance than Att SPP-nets.
- 2
- 2
- 2
- System dynamics modeling and simulation of exercise-based health promotion in the context of population aging1
The overall development of exercise-based health promotion must consider the interaction of government support, social support and individual support, and allocate resources rationally.
- 1
- 1
- 1
- 1
- 1
- 1
- 1
- Retracted on February 24, 2022 : SVR-ARMA Stock Price Prediction Model Based on Wavelet Noise Reduction1
ACM has decided to retract the Entire Conference Proceedings and all related papers from the ACM Digital Library and none of the papers from this Proceeding should be cited in the literature because of the questionable integrity of the peer review process for this Conference.
- 1
- 1
- Attentional Residual Dense Factorized Network for Real-Time Semantic Segmentation1
A residual dense factorized convolution block (RDFB) is designed, which reaps the benefits of low-level and high-level layer-wise features through dense connection to boost segmentation precision whilst enjoying efficient computation by factorizing large convolution kernel into the product of two smaller kernels.
- Stacked Marginal Time Warping for Temporal Alignment1
A stacked marginal time warping method is explored by using marginal stacked denoising autoencoder (mSDA) as the regularization term which enables SMTW to marginalize out noises and learn layer-wise non-linear representations with the effective closed-form solution.
- 1
- 1
- 1
- 1
- Research on traditional Chinese medicine CBR based on ontology1
Analysis of methods and technology used in CBR based on ontology from knowledge representation of clinical cases and similarity computation is made.
- 1
- Research on Psychology Data Clustering Algorithm Based on CUDA1
Aiming at the problem that traditional K-Means algorithm is sensitive to the initial clustering center, hierarchical clustering algorithm CURE is used to mitigate the problem and CUDA technology clustered several times, so as to improve the execution efficiency of the algorithm.
- MULTI-TEMPORAL AND MULTI-SENSOR IMAGE MATCHING BASED ON LOCAL FREQUENCY INFORMATION1
An efficient approach to automated multi-temporal and multi-sensor image matching based on local frequency information is presented and two new independent image representations, Local Average Phase (LAP) and Local Weighted Amplitude (LWA), are presented to emphasize the common scene information, while suppressing the non-common illumination and sensor-dependent information.
- 1
- Mechanical Product Concept Extraction Method1
Improved domain concept acquisition method of domain relevance and domain consensus, acquires machine concepts from requirements book of mechanical design domain automatically, to set up a complete mechanical knowledge expression model.
- Margin maximization feature weighting with better adaptability1
The better adaptive margin metric and the fuzzy weighting technique were introduced for developing the new optimal objective function and, as demonstrated by the authors' experimental studies, the proposed algorithm is more adaptive.
- 1
- 1
- 1
- Thinking Globally, Acting Locally: Merge Sort's Auto-Vectorization on Shuffle-Heavy ISAs–
AutoVecSort is proposed, a specialized auto-vectori-zation framework for merge sort that systematically addresses shuffle bottlenecks through a “Thinking Globally, Acting Locally” philosophy and outperforms highly optimized custom vectorized sorts, existing auto-vectorization frameworks, and widely used library implementations.
- –
- –
- –
- –
- A Kazakh–Chinese Cross-Lingual Joint Modeling Method for Question Understanding–
A Cross-lingual Token-level Bi-Interactive Model (Bi-XTM) is proposed, which introduces a novel subtask interaction method that leverages the token-level intent output distribution as additional information for slot vector representation, effectively reducing error propagation and enhancing the information exchange between intent and slot vectors.
- Feature decoupling and recombination for weakly-supervised semantic segmentation–
An explicit Feature Decoupling Module (FDM) is introduced in this paper, which strives to separate the features of co-occurring classes based on semantic aggregation and contrastive learning and proposes a Feature Recombination Module (FRM) that can recombine features of different classes to form new complete feature maps.
- –
- From Discrete to Continuous: Deep Fair Clustering With Transferable Representations–
This work designs an information bottleneck style objective function to learn fair and clustering-friendly representations and imposes fairness at the representation level, which could guarantee fairness for the transferred task regardless of clustering results.
- –
- –
- –
- –
- –
- –
- –
- –
- On the Equity of Nuclear Norm Maximization in Unsupervised Domain Adaptation–
A new property termed equity is identified, which indicates the balance degree of predicted classes, to demystify the efficacy of nuclear norm maximization for UDA theoretically, and offers a new discriminability-and-equity maximization paradigm built on squares loss, such that predictions are equalized explicitly.
- –
- Name-based Approach to Identify Suspicious Return Statements–
This work designs a sequence of heuristic rules to determine whether the return statement of a given method matches the signature of it, and proposes a name-based approach to identify suspicious return statements.
- Research on O2O Blending Teaching Application to Practical Course–
Advantages and disadvantages of pure online course are analyzed and one O2O blending teaching approach applied to practical course is proposed.
- –
- –
- –
- –
- –
- –
- –
- Semi-supervised Non-negative Local Coordinate Factorization–
Experimental results on UCI datasets and two popular face image datasets suggest that SNLCF outperforms the representative methods in terms of both average accuracy and average normalized mutual information.
- –
- –
- A Combined Method for Chinese Micro-Blogging Topic Tracking–
Experiments show that, compared with the model based on single vector model, classical TF-IDF and the tracking method of C.45Bagging similarity computing, this method have a better performance on precision, recall ratio and F1 value.
- –
- –
- –
- –
- –
- –
- –
- The new hybrid approach to protect MPEG-2 video header–
A new approach to detect MPEG-2 header errors using a MD5 hash function and information hiding technique is proposed and implemented and the results show that the new approach can detect 100% of header error at both block level and higher levels without increasing the transmission bit rate.
- –
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 2026-10-14. 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.