Is this you? Claim this profile to correct it, add a bio and choose the work people see first.
Claim this profileWorks281 from public data
- A survey of appearance models in visual object tracking771
This survey provides a detailed review of the existing 2D appearance models for visual object tracking and takes a module-based architecture that enables readers to easily grasp the key points ofVisual object tracking.
- Multi-View Representation Learning: A Survey from Shallow Methods to Deep Methods666
This survey aims to provide an insightful overview of theoretical foundation and state-of-the-art developments in the field of multi-view representation learning and to help researchers find the most appropriate tools for particular applications.
- Spectral clustering for multi-type relational data288
A general model, the collective factorization on related matrices, is proposed for multi-type relational data clustering and a novel algorithm is derived, the spectral relational clustering, to cluster multi- type interrelated data objects simultaneously.
- Deep Air Learning: Interpolation, Prediction, and Feature Analysis of Fine-Grained Air Quality274
The proposed approach utilizes the information pertaining to the unlabeled spatio-temporal data to improve the performance of the interpolation and the prediction, and performs feature selection and association analysis to reveal the main relevant features to the variation of the air quality.
- Spatio-Temporal Graph Routing for Skeleton-Based Action Recognition263
A novel spatio-temporal graph routing (STGR) scheme for skeletonbased action recognition, which adaptively learns the intrinsic high-order connectivity relationships for physicallyapart skeleton joints, is proposed.
- A General Model for Multiple View Unsupervised Learning244
The proposed model introduces the concept of mapping function to make the different patterns from different pattern spaces comparable and hence an optimal pattern can be learned from the multiple patterns of multiple representations.
- 234
- Co-clustering by block value decomposition197
A new co-clustering framework, block value decomposition(BVD), is presented, which factorizes the dyadic data matrix into three components, the row- coefficient matrix R, the block value matrix B, and the column-coefficient matrix C, which iteratively computes the three decomposition matrices based on the multiplicative updating rules.
- Episode-Based Prototype Generating Network for Zero-Shot Learning178
A novel generative model is proposed that synthesizes visual prototypes conditioned on the class semantic prototypes into an adversarial framework paired with a parameter-economic Multi-modal Cross-Entropy Loss to capture the discriminative information.
- Unsupervised learning on k-partite graphs160
A general model, the relation summary network, is proposed to find the hidden structures (the local cluster structures and the global community structures) from a k-partite graph to provide a principal framework for unsupervised learning on k- partite graphs of various structures.
- 158
- 153
- Incremental Tensor Subspace Learning and Its Applications to Foreground Segmentation and Tracking152
Experimental evaluations against state-of-the-art algorithms demonstrate the promise and effectiveness of the proposed incremental tensor subspace learning algorithm, and its applications to foreground segmentation and object tracking.
- 151
- Intelligent Indexing and Semantic Retrieval of Multimodal Documents149
This research addresses the problem of finding pictures by optimally combining text and image similarity in a MMIR system and presents a general model for multimodal information retrieval that addresses the following issues: users' information need, and determining the most appropriate weighted combination of indexing techniques in order to best satisfy information need.
- 140
- 138
- Improved prototypical networks for few-Shot learning132
This work proposes an attention-analogous strategy to explore the class distribution information by distributing different weights to samples based on their representativeness and proposes a distance scaling strategy to reduce the intra- class difference while enlarge the inter-class difference.
- 131
- 129
- 128
- 120
- 115
- 113
- 112
- Applying data mining in investigating money laundering crimes108
This work has identified a new paradigm of problems --- that of automatic community generation based on uni-party data, the data in which there is no direct or explicit link information available, and proposed a new methodology for Link Discovery based on Correlation Analysis (LDCA).
- Transductive Zero-Shot Learning With a Self-Training Dictionary Approach91
A bidirectional mapping-based semantic relationship modeling scheme that seeks for cross-modal knowledge transfer by simultaneously projecting the image features and label embeddings into a common latent space is proposed.
- Semisupervised Autoencoder for Sentiment Analysis90
The effectiveness of the usage of autoencoders in modeling textual data is investigated, and the model significantly outperforms all the competing methods with respect to classification accuracy and successfully learns highly discriminative feature maps, which explains its superior performance.
- Stacked Semantics-Guided Attention Model for Fine-Grained Zero-Shot Learning88
A novel stacked semantics-guided attention (S2GA) model to obtain semantic relevant features by using individual class semantic features to progressively guide the visual features to generate an attention map for weighting the importance of different local regions.
- S3Pool: Pooling with Stochastic Spatial Sampling87
This work proposes a novel pooling strategy with stochastic spatial sampling (S3Pool), where the regular downsampling is replaced by a more general stochastics version that acts as a strong regularizer and can be seen as doing implicit data augmentation by introducing distortions in the feature maps.
- 87
- Cross-media semantic representation via bi-directional learning to rank85
This paper proposes a general cross-media ranking algorithm to optimize the bi-directional listwise ranking loss with a latent space embedding, which it is called Bi- Directional Cross-Media Semantic Representation Model (Bi-CMSRM).
- Graph-Free Knowledge Distillation for Graph Neural Networks84
The proposed graph-free KD (GFKD) learns graph topology structures for knowledge transfer by modeling them with multinomial distribution and introduces a gradient estimator to optimize this framework.
- Linear discriminant analysis using rotational invariant L1 norm78
A novel rotational invariant L"1 norm (i.e., R"1norm) based discriminant criterion (referred to as DCL"1) is proposed, which better characterizes the intra-class compactness and the inter-class separability by using the rotations of the Frobenius norm.
- Zero-Shot Learning via Latent Space Encoding76
Extensive experimental results on four benchmark datasets clearly demonstrate the superiority of the proposed encoder–decoder approach, namely latent space encoding (LSE), on several ZSL tasks, including traditional ZSL, generalized Z SL, and zero-shot retrieval.
- Fine-tune BERT with Sparse Self-Attention Mechanism73
A novel Sparse Self-Attention Fine-tuning model (referred as SSAF) is developed which integrates sparsity into self-attention mechanism to enhance the fine-tuned performance of BERT.
- Cross-Relation Cross-Bag Attention for Distantly-Supervised Relation Extraction72
This paper proposes to conduct multi-instance learning with a novel Cross-relation Cross-bag Selective Attention (C2SA), which leads to noise-robust training for distant supervised relation extractor and employs the sentence-level selective attention to reduce the effect of noisy or mismatched sentences.
- META-DDIE: predicting drug–drug interaction events with few-shot learning70
A few-shot computational method named META-DDIE, which consists of a representation module and a comparing module, to predict DDI events and helps to identify the key factors that may cause DDI Events and reveal the relationship among different events.
- 69
- Deep Attentive Sentence Ordering Network68
A novel deep attentive sentence ordering network (referred as ATTOrderNet) which integrates self-attention mechanism with LSTMs in the encoding of input sentences enables us to capture global dependencies among sentences regardless of their input order and obtains a reliable representation of the sentence set.
- Multi-Relational Contrastive Learning Graph Neural Network for Drug-Drug Interaction Event Prediction67
A new method, Multi-Relational Contrastive learning Graph Neural Network, MRCGNN for brevity, to predict DDI events, which integrates the two aspects of information by deploying a GNN on the multi-relational DDI event graph attributed with the drug features extracted from drug molecular graphs.
- 67
- SBAT: Video Captioning with Sparse Boundary-Aware Transformer64
A novel method called sparse boundary-aware transformer (SBAT) is proposed to reduce the redundancy in video representation and an aligned cross-modal encoding scheme to boost the multimodal interaction.
- Deep Learning Driven Visual Path Prediction From a Single Image64
A deep learning framework is proposed, which simultaneously performs deep feature learning for visual representation in conjunction with spatiotemporal context modeling and a unified path-planning scheme is proposed to make accurate path prediction based on the analytic results returned by the deep context models.
- 64
- HimGNN: a novel hierarchical molecular graph representation learning framework for property prediction61
This work proposes a novel framework for molecular property prediction, called hierarchical molecular graph neural networks (HimGNN), and proposes a simple yet effective rescaling module, called contextual self-rescaling, that adaptively recalibrates molecular representations by explicitly modelling interdependencies between atom and motif features.
- GNAS61
This work proposes an efficient greedy neural architecture search approach (GNAS) to automatically discover the optimal tree-like deep architecture for multi-attribute learning and demonstrates the effectiveness and compactness of neural architectures derived by GNAS.
- Deep learning driven blockwise moving object detection with binary scene modeling60
A deep learning based block-wise scene analysis method equipped with a binary spatio-temporal scene model based on the stacked denoising autoencoder that ensures the high efficiency of moving object detection.
- 58
- TVT: Two-View Transformer Network for Video Captioning.57
A novel video captioning framework, i.e., Two-View Transformer (TVT), which comprises of a backbone of Transformer network for sequential representation and two types of fusion blocks in decoder layers for combining different modalities effectively.
- 57
- Obstacle detection based on qualitative and quantitative 3D reconstruction55
Three different algorithms for mobile robot vision-based obstacle detection are presented in this paper each based on different assumptions and it is concluded that in terms of the robustness of performance, the third algorithm is superior to the other two.
- 53
- A low rank structural large margin method for cross-modal ranking52
A general cross-modal ranking algorithm to optimize the listwise ranking loss with a low rank embedding, which is called Latent Semantic Cross-Modal Ranking (LSCMR) and shows significant improvements over the state-of-the-art methods.
- Text Guided Person Image Synthesis51
It shows for the first time that the person image can automatically edit from the natural language descriptions, and synthesizes a realistic and appearance transferred person image according to the text in conjunction with the target pose.
- 51
- 49
- RKOF: Robust Kernel-Based Local Outlier Detection47
This work adopts the variable kernel density estimate and the weighted neighborhood density estimate to improve the robustness to the variations of the parameter k, while keeping the same framework with LOF, and proposes a novel kernel function named the Volcano kernel, which is more suitable for outlier detection.
- Geometric histogram: a distribution of geometric configurations of color subsets47
Experimental results demonstrate that triangular geometric histogram, one of the simplest special cases of geometric histograms, is more desirable than existing techniques for content-based image retrieval, especially when the database in question consists of on-line color images which are extremely heterogenous in terms of the content of images, camera types, lighting conditions and so on.
- Joint structured pruning and dense knowledge distillation for efficient transformer model compression45
A novel Joint Model Compression method by combining structured pruning and dense knowledge distillation techniques to significantly compress original large language model into a deep compressed shallow network to achieve a significant compression without sacrificing model accuracy.
- 45
- 43
- Robust foreground segmentation based on two effective background models42
Two block-based background models are proposed using the recently developed incremental rank-(R1, R2, R3) tensor-based subspace learning algorithm to fully capture the intrinsic spatio-temporal characteristics of a scene, leading to robust foreground segmentation results.
- 41
- Text Coherence Analysis Based on Deep Neural Network39
A novel deep coherence model (DCM) using a convolutional neural network architecture to capture the text coherence and can be easily trained in an end-to-end fashion.
- BERT-enhanced Relational Sentence Ordering Network38
A novel BERT-enhanced Relational Sentence Ordering Network (referred to as BRSON) is introduced by leveraging BERT for capturing better dependency relationship among sentences to enhance the coherence modeling for the entire paragraph.
- Multimedia Data Mining38
This novel, self-contained book examines how the merging of multimedia and data mining research can promote the understanding and advance the development of knowledge discovery in multimedia data.
- Manifold regularized cross-modal embedding for zero-shot learning37
A Manifold regularized Cross-Modal Embedding (MCME) approach for ZSL is proposed by formulating the manifold constraint for intrinsic structure of the visual features as well as aligning pairwise consistency in order to preserve the locally visual structure in the embedding process.
- Dropout Training of Matrix Factorization and Autoencoder for Link Prediction in Sparse Graphs37
This paper applies dropout to training both the MF and AE parts, and shows that it can significantly prevent overfitting by acting as an adaptive regularization, and proposes MF+AE: a model training MF andAE jointly with shared parameters.
- 37
- 36
- 35
- Scientific articles recommendation33
A novel matrix factorization model, the topic regression Matrix Factorization (tr-MF), that is effective in making predictions for users with only few ratings or even no ratings, and supports tasks that are specific to a certain field.
- Relational clustering by symmetric convex coding33
A general model for relational clustering based on symmetric convex coding is proposed that is applicable to all types of relational data and unifies the existing graph partitioning formulation.
- BALAS: Empirical Bayesian learning in the relevance feedback for image retrieval32
The experimental evaluations have confirmed the rationale of the proposed ranking scheme, and have demonstrated that BALAS is superior to an existing relevance feedback method in the current literature in capturing the overall retrieval semantics.
- 30
- 30
- 30
- Scalable Nonnegative Matrix Factorization with Block-wise Updates29
By leveraging a new form of update functions, this paper can perform local aggregation and fully explore parallelism, and perform frequent updates, which aim to use the most recently updated data whenever possible, and are more efficient than their traditional concurrent counterparts.
- Dual Low-Rank Multimodal Fusion28
Empirical studies on benchmark multimodal analysis tasks show that the proposed methods outperform the state-of-the-art tensor-based fusion methods with a similar computational complexity.
- Recurrent convolutional video captioning with global and local attention27
This paper introduces a novel video captioning method, Recurrent Convolutional Video Captioning with Global and Local Attention (RCGL), and demonstrates that RCGL outperforms the state-of-the-art in four common metrics.
- Complementary Calibration: Boosting General Continual Learning With Collaborative Distillation and Self-Supervision26
A new collaborative distillation approach for addressing the relation deviation is proposed by utilizing ensemble dark knowledge of new model’s outputs and reserved outputs, which maintains the performance of old tasks as well as balancing the relationship among all classes.
- Addressing CBIR efficiency, effectiveness, and retrieval subjectivity simultaneously26
Theoretical analysis and experimental evaluations show that FAST methodology holds a great promise in delivering fast and semantics-tailored image retrieval in CBIR.
- 25
- Multi-modal generative adversarial network for zero-shot learning25
This paper proposes a novel approach for Zero-Shot Learning (ZSL), where the test instances are from the novel categories that no visual data are available during training, and fuse various types of class semantic prototypes resided in different semantic spaces with a feature fusion network to generate pseudo visual features.
- A latent topic model for linked documents25
This paper proposes a citation-topic (CT) model which assumes a probabilistic generative process for corpora and captures the transitivity of the relations among documents.
- 25
- Reweighting and information-guidance networks for Few-Shot Learning24
Two strategies on the basis of Prototypical Networks are proposed to improve the discriminativeness and representativeness of the visual prototypes for few-shot learning task and could be seamlessly combined into some existing approaches to substantially improve their performances on few- shot classification.
- 24
- 24
- Discriminative feature selection for multi-view cross-domain learning24
Theoretic analysis and extensive evaluations in the specific application of object identification and image classification against several state-of-the-art methods demonstrate the outstanding superiority of DISMUTE, based on discriminative feature selection for multi-view cross-domain learning.
- Mining partially annotated images24
The Semi-Supervised Correspondence Hierarchical Dirichlet Process (SSCHDP) is proposed, which lies in exploiting the information pertaining to partially annotated images or even unannotated images to achieve semi-supervised learning under the HDP structure.
- Automatic medical image annotation and retrieval24
This paper presents a novel medical image retrieval method for a special medical image retrieved problem where the images in the retrieval database can be annotated into one of the pre-defined labels.
- 24
- Multi-Task Learning with Gaussian Matrix Generalized Inverse Gaussian Model23
This paper proposes the GMGIG regression model for multi-task learning, which is superior to the peer methods in regression and prediction, and proposes two sampling strategies for computing the statistics of the MGIG distribution.
- The heterogeneous feature selection with structural sparsity for multimedia annotation and hashing: a survey23
This paper introduces many of the recent efforts in sparsity-based heterogenous feature selection, the representation of the intrinsic latent structure embedded in multimedia, and the related hashing index techniques.
- 23
- Learning with Retrospection22
Learning with retrospection (LWR) is a simple yet effective training framework to improve accuracies, calibration, and robustness of DNNs without introducing any additional network parameters or inference cost, but only with a negligible training overhead.
- Low-Rank HOCA: Efficient High-Order Cross-Modal Attention for Video Captioning22
This paper proposes a video captioning model with High-Order Cross-Modal Attention (HOCA) where the attention weights are calculated based on the high-order correlation tensor to capture the frame-level cross-modal interaction of different modalities sufficiently and introduces Low-Rank HOCA which adopts tensor decomposition to reduce the extremely large space requirement of H OCA.
- LSTM-in-LSTM for generating long descriptions of images22
This model outperforms state-of-theart methods on several benchmark datasets when used to generate long rich fine-grained descriptions of given images in terms of four different metrics (BLEU, CIDEr, ROUGE-L, and METEOR).
- A general framework for relation graph clustering22
Under this framework, a family of clustering algorithms including both hard and soft versions, which are capable of learning cluster patterns from relation graphs with various structures and statistical properties are derived.
- Coordinating Experience Replay: A Harmonious Experience Retention approach for Continual Learning21
A Coordinating Experience Replay approach consisting of a Harmonious Experience Retention approach and a Vital Exemplar Reserved Sampling strategy, which constrains the rehearsal process and develops two criteria to select examples for better replaying the beneficial experience is proposed.
- Recognition of adult images, videos, and web page bags21
An integrated adult-content recognition system which can detect adult images, adult videos, and adult Web page bags and model skin patches rather than skin pixels, resulting in better results than state-of-the-art algorithms which model skin pixels.
- 21
- A Holistic, In-Compression Approach to Video Segmentation for Independent Motion Detection21
A holistic, in-compression approach to efficient video segmentation that is close to or even faster than real-time in "normal" platforms while still maintaining a good quality segmentation.
- 20
- 20
- Joint Multilabel Classification With Community-Aware Label Graph Learning20
This work proposes a multilabel classification framework based on a joint learning approach called label graph learning (LGL) driven weighted Support Vector Machine (SVM) which explicitly models the inter-label correlations by LGL, which is jointly optimized with multILabel classification in a unified learning scheme.
- Semi-supervised Learning Based on Semiparametric Regularization20
This paper addresses the semi-supervised learning problem by developing a semiparametric regularization based approach, which attempts to discover the marginal distribution of the data to learn the parametric function through exploiting the geometric distribution ofThe data.
- Tolerant Self-Distillation for image classification19
Two metrics from the intra-class distribution of correct-predicted and incorrect-predicted samples are introduced to provide a new perspective on the overfitting issue and a knowledge distillation approach without pretraining a teacher model in advance named Tolerant Self-Distillation (TSD) is proposed for alleviating the overfitting issue.
- Zero-shot classification with unseen prototype learning19
This paper proposes a novel Unseen Prototype Learning (UPL) model, which is a simple yet effective framework to learn visual prototypes for unseen categories from the corresponding class-level semantic information, and the learned features are treated as latent classifiers directly.
- 19
- Enhanced max margin learning on multimodal data mining in a multimedia database19
A new max margin learning approach called Enhanced Max Margin Learning (EMML) framework is developed and applied to the Berkeley Drosophila embryo image database, and the performance comparison with a state-of-the-art multimodal data mining method is reported.
- Structural Correspondence Learning for Cross-Lingual Sentiment Classification with One-to-Many Mappings18
This paper proposes a cross-lingual SCL based on distributed representation of words; it can learn meaningful one-to-many mappings for pivot words using large amounts of monolingual data and a small dictionary and the experimental results show that this approach is more competitive than the state-of-the-art methods in cross-lingsual sentiment classification.
- Generative Models for Evolutionary Clustering18
Two generative models are developed that substantially advance the literature on evolutionary clustering, in the sense that not only do they both perform better than those in the existing literature, but more importantly, they are capable of automatically learning the cluster numbers and explicitly addressing the corresponding issues.
- Improving Web search using image snippets18
It is shown that Web search performance can be significantly improved if imagery is considered in interactive Web search, and the promise of introducing multimodal techniques into the proposed interactive Websearch framework is demonstrated.
- 18
- Joint categorization of queries and clips for web-based video search17
Evaluation of the developed system and user study has indicated that the joint categorization of queries and video data boosts the video search relevance and user search experience.
- Stretching Bayesian Learning in the Relevance Feedback of Image Retrieval17
An integrated ranking scheme in BALAS is developed which complementarily combines the subjective relevancy confidence and the objective feature-based distance measure to capture the overall retrieval semantics.
- 16
- Sparsity-control ternary weight networks16
This paper proposes the first sparsity-control approach (SCA) to train ternary weight networks, which is simply achieved by a weight discretization regularizer (WDR).
- 16
- 16
- 16
- 16
- 15
- 15
- Multi-Channel Pyramid Person Matching Network for Person Re-Identification15
A Multi-Channel deep convolutional Pyramid Person Matching Network (MC-PPMN) based on the combination of the semantic-components and the color-texture distributions to address the problem of person re-identification is presented.
- Scalable Linear Visual Feature Learning via Online Parallel Nonnegative Matrix Factorization15
A parallel NMF method called alternating least square block decomposition (ALSD), which efficiently solves a set of conditionally independent optimization subproblems based on a highly parallelized fine-grained grid-based blockwise matrix decomposition, can be effectively implemented in a MapReduce-based Hadoop framework.
- 15
- Robust Visual Tracking Based on an Effective Appearance Model15
Theoretic analysis and experimental evaluations demonstrate the promise and effectiveness of the proposed novel spatial Log-Euclidean appearance model (SLAM), capable of capturing both the global and local spatial layout information of object appearance by constructing a block-based Log- eigenspace representation.
- Zero-shot learning with Multi-Battery Factor Analysis14
This paper develops an algorithm named Multi-Battery Factor Analysis (MBFA) to build a unified semantic space, and then employs multiple types of side information in it to achieve the ZSL.
- Structured Embedding via Pairwise Relations and Long-Range Interactions in Knowledge Base14
This paper introduces Path-Ranking to capture the long-range interactions of knowledge graph and at the same time preserve the pairwise relations of knowledgeGraph (SePLi), which achieves better performances of embeddings.
- 14
- Formation of intercalation compound of kaolinite–glycine via displacing guest water by glycine14
Results show that glycine can only be intercalated into hydrated kaolinite to form glycine-kaolinite by utilizing water molecules as a transition phase.
- 14
- 13
- 13
- 13
- Relational Data Clustering: Models, Algorithms, and Applications12
The field is defined, different types of model formulations for relational data clustering are introduced, various algorithms for the corresponding models are presented, and examples of how to apply these models and algorithms to solve real-world problems are shown.
- Relational Data Clustering12
The field is defined, different types of model formulations for relational data clustering are introduced, various algorithms for the corresponding models are presented, and examples of how to apply these models and algorithms to solve real-world problems are shown.
- 11
- Multimedia indexing and retrieval11
There was a paper session focusing on various topics in image retrieval, as well as models for search involving several media, and an invited talk given by Donna Harman of NIST about building a video database that will serve as a testbed for evaluating algorithms.
- 10
- 10
- 10
- 10
- Sketch the storyline with CHARCOAL: a non-parametric approach10
The adaptation of non-parametric nature in CHARCOAL makes the model can adaptively learn the appropriate number of stories and topics from news corpus and demonstrates both interpretability and superiority of the proposed approach.
- 10
- 10
- 10
- 10
- 10
- A data mining approach to modeling relationships among categories in image collection10
The experimental evaluations have demonstrated that the proposed approach models the semantic relationships effectively and the image retrieval prototype system utilizing the derived model is promising both in effectiveness and efficiency.
- Finding pictures in context10
This research focuses on improving precision and recall in a MMIR system by interactively combining text processing with image processing (IP) in both the indexing and retrieval phases.
- 10
- Uncertainty-aware enhanced dark experience replay for continual learning9
A novel sam- pling strategy dubbed Uncertainty-Aware Sampling (UAS) strategy, which employs model and data uncertainties as criteria to select samples that are stable to the model and have low noise for rehearsal, and a dual network to acquire new knowledge while maintaining old knowledge is designed.
- 9
- 9
- 9
- Learning with Marginalized Corrupted Features and Labels Together9
This paper proposes Regularized Marginalized Cross-View learning (RMCV) by jointly modeling on attribute noise and label noise and demonstrates that RMCV outstands with a superior performance in comparison with state-of-the-art methods.
- 9
- Improve web search using image snippets9
This paper shows that the Web search performance can be enhanced if image information is considered and a new Web search framework is proposed, where image snippets are extracted for the Web pages, which are then provided along with text snippets to the user such that it is much easier and more accurate for the user to identify the Webpages he or she expects and to reformulate the initial query.
- 9
- Exploiting Multimodal Context in Image Retrieval9
This research focuses on improving precision and recall in an MMIR system by interactively combining text processing with image processing (IP) in both the indexing and retrieval phases.
- Multi-Content Interaction Network for Few-Shot Segmentation8
A Multi-Content Interaction Network (MCINet) is proposed to remedy FSS issues by fully exploiting and interacting with the different contextual information contained in distinct branches by boosting the query representations and enhancing the support-query correlations.
- 8
- 8
- Learning with limited and noisy tagging8
A discriminative model is proposed, called SpSVM-MC, that exploits both labeled and unlabeled data through a semi-parametric regularization and takes advantage of the multi-label constraints into the optimization of a robust tagging method.
- Multi-view learning from imperfect tagging8
A novel discriminative classification method is proposed under the framework of MITL, which explicitly makes use of the given multiple labels simultaneously as an additional feature to deliver a more effective classification performance than the existing literature where one label is considered at a time as the classification target.
- Mining noisy tagging from multi-label space8
It is argued that all the given multiple tags can be utilized simultaneously as an additional feature and the information contained in the multi-label space can be taken advantage of to improve the performance of the classification.
- Unsupervised Ensemble Learning for Mining Top-n Outliers8
A general framework of ensemble learning in the top-n outlier detection is proposed based on the rank aggregation techniques and shows a stable and effective performance independent of different datasets in a good scalability in comparison with the state-of-the-art literature.
- 8
- Using data mining techniques for building fusion models8
This paper will provide the details for the motivation, theory and techniques which the author has developed, as well as the results of a set of experiments.
- Three-dimensional reconstruction under varying constraints on camera geometry for robotic navigation scenarios8
3D reconstruction techniques developed under special constraints on camera geometry have been applied to a robotic navigation scenario and it is shown that given four coplanar correspondences, 3D structures can be recovered up to two solutions and with only one uniform scale factor.
- Multi-Proxy Learning from an Entropy Optimization Perspective7
This work presents an easy-to-implement framework to effectively capture the local neighbor relationships via learning multiple proxies for each class that collectively approximate the intra-class distribution under the multi-proxy learning framework and provides a training routine that both minimizes the entropy of intra- class probability distribution and maximizesThe entropy of inter-class probability distribution.
- Full-Resolution Correspondence Learning for Image Translation7
The proposed GRU-assisted PatchMatch is fully differentiable and highly efficient, and performs considerably better than state-of-the-arts on producing high-resolution images.
- Learning a Key-Value Memory Co-Attention Matching Network for Person Re-Identification7
The proposed KVM-MN model is capable of building an effective local-position-aware person representation that encodes the spatial feature information in the form of multi-head key-value memory.
- Attention Based Recurrent Neural Networks for Online Advertising7
This work investigates the use of recurrent neural networks in the context of online advertising, where RNNs are used to map both query and ads to real valued vectors, and proposes an attention network that assigns scores to different word locations according to their intent importance.
- Multimodal Data Mining in a Multimedia Database Based on Structured Max Margin Learning7
This article exploits the relations among different modalities in a multimedia database and presents a framework for general multimodal data mining problem where image annotation and image retrieval are considered as the special cases and proposes a new max margin structure learning approach called Enhanced Max Margin Learning (EMML) framework.
- Tracking news article evolution by dense subgraph learning7
This work proposes a context-dependent news knowledge discovery method based on temporally successive news article connection using subgraph learning that is able to adaptively construct a cross-article link network along the temporal dimension, and effectively discovers the news event pattern by dense sub graph learning using the contextual news connection structures.
- Visual search reranking with RElevant Local Discriminant Analysis7
The proposed RELDA algorithm preserves the local manifold structure of the whole data as well as controls the relevance between labeled examples and has an analytic form of the globally optimal solution and can be computed based on eigen-decomposition.
- KDD/MDM 20067
The central theme of this year's MDM workshop is to merge multimedia and data mining research together to exploit the synergy between the two areas to advance and promote the research and development in multimedia data mining.
- 7
- 7
- Image Annotation Retrieval with Text-Domain Label Denoising6
This work proposes a technique for calculating a set of probabilities for the relevance of each tag for a given image relying soley on information in the text domain, namely through widely-available pretrained continous word embeddings.
- 6
- 6
- 6
- Learning Bregman Distance Functions for Structural Learning to Rank6
The proposed model Robust Structural Bregman distance functions Learning to Rank (RSBLR) is a general and unified framework for learning distance functions to rank and shows the superiority of this method to the state-of-the-art literature, as well as its robustness to the noisily labeled outliers.
- Discriminative Transfer Learning on Manifold6
A discriminative regression model is imposed over the latent factors to enhance the capability of label prediction and to minimize the Maximum Mean Discrepancy in the latent manifold subspace, as opposed to typically in the original data space, to bridge the gap between different domains.
- Robust object tracking using a spatial pyramid heat kernel structural information representation6
This paper takes advantage of heat kernel structural information (HKSI) matrices to represent object appearance, because HKSI matrices perform well in characterizing the edge flow information on the object appearance graph.
- Editorial: Introduction to the Special Issue on Multimedia Data Mining6
This special issue evolved from a successful workshop organized in conjunction with the 2006 ACM KDD conference, but the special issue was open to the whole community.
- 6
- 6
- 5
- Balancing Feature Alignment and Uniformity for Few-Shot Classification5
This paper proposes a unified framework that perturbs the feature embedding space using two low-bias estimators and effectively combines knowledge distillation between class-wise pairs and enlarges the diversity in feature representations.
- Personalized Education: Blind Knowledge Distillation5
Personalized education (PE) is proposed to first help each student adaptively find its own blind knowledge region (BKR) where the student has not captured the knowledge from the teacher, and then teach the student on this region.
- COREN: Multi-Modal Co-Occurrence Transformer Reasoning Network for Image-Text Retrieval5
This paper proposes a unified multi-modal Co-Occurrence transformer Reasoning Network, dubbed as COREN, to comprehensively discover the semantic correlations of the two modalities and proposes a novel Adaptive Similarity Aggregation (ASA) module to achieve a more accurate cross- modal similarity measurement.
- 5
- An Embarrassingly Simple Approach to Training Ternary Weight Networks.5
ESA is the first sparsity-controlling approach to training ternary weight networks and beats the state-of-the-art approaches significantly and matches the performances of the full precision weight networks.
- Relative Attribute Learning with Deep Attentive Cross-image Representation5
A deep attentive cross-image representation learning (DACRL) model is introduced, which first extracts single- image representation with one shared subnetwork, and then learns attentiveCrossimage representation through considering the channel-wise attention of concatenated single-image feature maps.
- Boosted Zero-Shot Learning with Semantic Correlation Regularization5
This work proposes a unified framework, Boosted Zero-shot classification with Semantic Correlation Regularization (BZ-SCR), capable of capturing both discriminative and adaptable feature-to-class semantic alignments, while ensuring the reliability and adaptability of the learned samples.
- Multi-view learning with limited and noisy tagging5
This paper proposes a discriminative model, called MSMC, that exploits both labeled and unlabeled data through a semi-parametric regularization and takes advantage of the multi-label space consistency into the optimization.
- Independent motion detection directly from compressed surveillance video5
A novel solution to the problem frequently encountered in military/intelligence surveillance: automatic mining video data for independently moving targets based on the Linear System Consistency Analysis, and thus is called LSCA.
- 4
- Learning with Feature Network and Label Network Simultaneously4
Extensive evaluations on three benchmark data sets demonstrate that DRML outstands with a superior performance in comparison with some existing multi-label learning methods.
- Pattern change discovery between high dimensional data sets4
The notion of the principal angles between the subspaces is introduced to measure the subspace difference between two high-dimensional data sets and it is shown how matrix factorization can be naturally embedded into the likelihood ratio test based on the linear models.
- A Topic Model for Linked Documents and Update Rules for its Estimation4
A Bernoulli Process Topic (BPT) model is presented, where the interdependence among the documents is modeled by a randomBernoulli process, and it could be applied to many applications including detecting the topics from corpora and clustering the documents.
- 4
- Subspace morphing theory for appearance based object identification4
Theoretical analysis and experimental evaluation show that in the situation where images are provided in different dimensions, subspace morphing theory is superior to the existing, normalization-based techniques in performance.
- Show A Tell: Using Speech Input for Image Interpretation and Annotation*4
This research concerns the exploitation of linguistic context in vision and proposes new techniques for using qualitative contextual information, but also efficiently exploits existing image interpretation technology.
- Multi-Scale Similarity Aggregation for Dynamic Metric Learning3
A new multi-scale similarity aggregation method (MSA) for dynamic metric learning (DyML), which adopts a pretraining-finetuning scheme and efficiently learns the similarity relationship for each semantic level.
- 3
- Multi-label Triplet Embeddings for Image Annotation from User-Generated Tags3
This work studies the representational embedding of images and their corresponding annotations--in the form of tag metadata--such that, given a piece of the raw data in one modality, the corresponding semantic description can be retrieved in terms of theRaw data in another.
- Learning With Incomplete Labels3
This paper introduces the Improved Cross-View learning with Outlier Detection (referred as ICVL-OD) model, which considers both global and local patterns of label relationship to enrich the original label set and outstand with superior performances in comparison with the competing methods.
- 3
- Distributed cross-media multiple binary subspace learning3
A semi-supervised learning method, named Multiple Binary Subspace Regression, for cross-media data concept detection, which project the original cross- media data onto the same subspace-level representation simultaneously by mapping to the corresponding subspaces for dimensionality reduction.
- Scientific articles recommendation with topic regression and relational matrix factorization3
This paper presents a novel topic regression model, the topic regression matrix factorization (tr-MF), to solve the problem of recommending scientific articles to users in an online community with a new perspective of considering topic regression modeling and articles relational structure analysis simultaneously.
- 3
- Spectral Graph Partitioning Based on a Random Walk Diffusion Similarity Measure3
A novel random walk diffusion similarity measure (RWDSM) is proposed and a novel spectral graph partitioning algorithm (referred to as DGPCA) is developed, designed to effectively characterize the intra-class compactness and the inter-class separability.
- Mining Bulletin Board Systems Using Community Generation3
A uni-party data community generation algorithm called ISGI is proposed, which employs a specifically designed hierarchical similarity function to measure the correlations between two different individual users.
- 3
- Comparing various algorithms for discovering social groups with uni-party data3
This paper introduces several new algorithms and analyzes their ability to generate a prioritized ranking of individuals involved in the Ponzi scheme based on their individual activity to compare the accuracy of each algorithm.
- Meta-Transfer Networks for Zero-Shot Learning2
This work proposes an episode-based approach to accumulate experiences on addressing disjoint-ness issue by mimicking extensive classification scenarios where training classes and test classes are disjoint.
- Efficient Network Representations Learning: An Edge-Centric Perspective2
An edge-centric view where features are associated with edges, in addition to the network structure, is provided, which provides a fine-grained characterization of a network, where the dynamic interactions between vertices are considered instead of the static property of vertices.
- 2
- 2
- A Multiple-Instance Learning Based Approach to Multimodal Data Mining2
This approach is a highly scalable and adaptable framework that the authors call co-learning and is applied to the Berkeley Drosophila ISH embryo image database for the evaluations of the mining performance in comparison with a state-of-the-art multimodal data mining method.
- Guest Editors' introduction to the special issue: machine learning approaches to multimedia information retrieval2
Effectiveness refers to how well the retrieved multimedia information meets the users’ expectation, and the efficiency addresses how responsive the underlying retrieval mechanisms are as queries are posed by the users.
- Discovering social groups without having relational data2
This paper presents one class of problems which is called the uni-party data community generation paradigm, and discusses various applications, a methodology and results from two problem domains.
- 2
- 1
- 1
- 1
- 1
- 1
- 1
- FR-ANet: A Face Recognition Guided Facial Attribute Classification Network1
A Face Recognition guided facial Attribute classification Network, called FR-ANet, where all the attributes share low-level features, while high- level features are specially learned for attribute groups.
- 1
- 1
- 1
- 1
- Societally connected multimedia across cultures1
This position paper proposes to define a new research area called ‘connected multimedia’, which is the study of a collection of research issues of the super-area social media that receive little attention in the literature.
- Guest Editors' Introduction: Special Section on Connected Multimedia1
This editorial is co-authored by the guest-editors who also organized the two editions of the workshop on the newly emerged theme of connected multimedia held in Hangzhou, China and in Florence, Italy, in October, 2010.
- Overview of ACM international workshop on connected multimedia1
In this workshop overview, what is meant by connected multimedia is defined, and the program of this workshop is briefly overviewed.
- Machine Learning Approaches to Link-Based Clustering1
This chapter reviewed several state-of-the-art machine learning approaches to different types of link-based clustering, including the spectral clustering for heterogeneous relational data, the symmetric convex coding for homogeneous relationalData, the citation model for clustering the special but popular homogeneous database data, and the statistical graphical model for dynamic relational clustering.
- Solving the Small and Asymmetric Sampling Problem in the Context of Image Retrieval1
The experimental evaluations have confirmed the rationale of the proposed ranking scheme, and have demonstrated that BALAS is superior to an existing relevance feedback method in the current literature in capturing the overall retrieval semantics.
- 1
- 1
- 1
- 1
- Querying non‐uniform image databases for biometrics‐related identification applications1
This research project focuses on developing techniques and technologies for automatically identifying human faces from images in the situations where face sample collections in the database as well as in the input query images are “as is”, i.e. no standard data collection environment is available.
- 1
- 1
- 1
- Using Region Analysis Technique to Efficiently Retrieve Online Image Databases1
This research investigates the image retrieval problem for online image databases by proposing a novel solution, called CLEAR, that incorporates multi-features including color, texture, shape, as well as conventional geometric information, which allows the features to be more descriptive in indexing the objects of an image.
- –
- –
- –
- –
- Marginal Fisher Regression Classification for Face Recognition–
The MFRC aims at minimizing the within-class compactness over the between-class separability to find an optimal embedding matrix for the LRC so that the L RC on that subspace achieves a high discrimination for classification.
- –
- –
- –
- Multilayer Semantic Analysis in Image Databases–
The large-scale extensive experimental results show that the proposed higher-level visual representation outperforms the traditional part-based image representations in retrieval, classification, and object recognition.
- A Highly Scalable and Adaptable Co-Learning Framework on Multimodal Data Mining in a Multimedia Database–
This chapter presents a highly scalable and adaptable co-learning framework on multimodal data mining in a multimedia database based on the multiple instance learning theory, which enjoys a strong scalability in the sense that the query time complexity is a constant, independent of the database scale.
- Knowledge discovery from relations–
The focus of this dissertation is on developing systematic machine learning approaches for different relational data available in various data mining tasks including supervised learning, unsupervised learning, and semi- supervised learning.
- A Holistic, In-Compression Approach to Mining Independent Motion Segments for Massive Surveillance Video Collections–
This chapter describes a large scale surveillance video data mining approach for those segments that contain independently moving targets, called Linear System Consistency Analysis (LSCA), to efficient video datamining for those independent motion segments.
- –
- –
- Introduction to the Eighth International Workshop on Multimedia Data Mining–
MDM 2007 contains six regular papers focusing on various aspects of multimedia data mining, including content-supported learning of user preferences, integration of multiple visual learners for media mining, salient-point based image object detection, the use of relevance feedback for video event mining, association rule mining for infrequent database items, and a novel algebra for supporting ranked data processing in multimedia mining applications.
- On Developing Theory and Application of Community Generation–
The spectral relational clustering algorithm the authors have developed is a powerful relational data clustersering algorithm that can be used for any type of relational data clustering and the general relation summary network model is the most general model.
- Developing machine learning techniques for real world applications–
It is proved in theory that a semi-supervised learning algorithm under this framework is probabilistically guaranteed to have the accuracy increased when the number of iterations increases, and it is proved that the application oriented accuracy is maximized at each iteration of the learning procedure.
- Real-Time Decision Support for Course of Action/Enemy Course of Action (COA/ECOA) Analysis–
This research has developed a specific solution as a method to identify whether a specific pair of commander's intent and COA is in alignment and if not, how far they divert from each other.
- –
- Qualitive Detection of Independently Moving Targets in MPEG Video Within the Compressed Domain–
A holistic, in-compression approach to efficient video prostanding that is close to or even faster than real-time in "normal" platforms while still maintaining comparable quality with state-of-the-art methods.
- A multimodal approach to time-invariant scene retrieval from single overhead image–
A multimodal approach is developed to solving the practical problem of time-invariant scene retrieval from single airport overhead image, in which an efficient method using image modality data is developed for time-specific object detection, and a method using graphics modalityData for removal of the detected objects.
- KNOWLEDGE SUPERVISED PERCEPTUAL GROUPING BASED QUALITATIVE BUILDING DETECTION FROM MONOCULAR AERIAL IMAGES–
This paper addresses an important and practical problem in computer vision and pattern recognition — qualitative target detection from aerial images and discusses the problem of qualitatively target detectionfrom aerial images.
- –
- –
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.