Is this you? Claim this profile to correct it, add a bio and choose the work people see first.
Claim this profileWorks455 from public data
- 226
- Fuzzy reasoning spiking neural P system for fault diagnosis218
This work extends SN P systems by introducing some new ingredients (such as three types of neurons, fuzzy logic and new firing mechanism) and proposes the fuzzy reasoning spiking neural P systems (FRSN P systems), which are particularly suitable to model fuzzy production rules in a fuzzy diagnosis knowledge base and their reasoning process.
- Graph-to-Tree Learning for Solving Math Word Problems213
Graph2Tree is proposed, a novel deep learning architecture that combines the merits of the graph-based encoder and tree-based decoder to generate better solution expressions for math word problem (MWP) solution expressions.
- 185
- Learning Sparse Representation With Variational Auto-Encoder for Anomaly Detection185
A novel sparse representation framework that learns dictionaries based on the latent space of variational auto-encoder for large-scale data sets and outperforms competing algorithms in all kinds of anomaly detection tasks.
- 183
- Zero-Shot Hashing via Transferring Supervised Knowledge162
A novel hashing scheme, termed zero-shot hashing (ZSH), which compresses images of "unseen" categories to binary codes with hash functions learned from limited training data of " seen" categories is proposed and an efficient alternating algorithm is developed to solve the ZSH model.
- 150
- Impulsive synchronization of coupled delayed neural networks with actuator saturation and its application to image encryption132
In this paper, impulsive synchronization of coupled delayed neural networks with actuator saturation is presented and the proposed image encryption system has high security properties.
- Learning in high-dimensional multimedia data: the state of the art129
A survey of feature transformation, feature selection and feature encoding, three approaches fighting the consequences of the curse of dimensionality on multimedia learning.
- 101
- Modeling of nonlinear system based on deep learning framework99
It is shown by experimental results that the proposed model with fewer parameters and iteration number can reconstruct the nonlinear system accurately and depict the memory effect, the non linear distortion and the dynamic performance of system precisely.
- Mining Contextual Information Beyond Image for Semantic Segmentation98
This paper proposes to mine the contextual information beyond individual images to further augment the pixel representations and proposes a representation consistent learning strategy to make the classification head better address intra-class compactness and inter-class dispersion.
- UQLIPS: a real-time near-duplicate video clip detection system98
Using a large collection of TV commercials, UQLIPS clearly demonstrates that it is feasible to perform real-time NDVC detection with high accuracy.
- Abnormal event detection for video surveillance using deep one-class learning88
A novel end-to-end model which integrates the one-class Support Vector Machine (SVM) into Convolutional Neural Network (CNN), named Deep One-Class (DOC) model is proposed, and the robust loss function derived from theOne-class SVM is proposed to optimize the parameters of this model.
- An anomaly-introduced learning method for abnormal event detection84
This work proposes an anomaly-introduced learning (AL) method to detect abnormal events, which makes use of abnormal information and prune testing instances with a coarse filter and reduce time cost of computing SRC.
- Heterogeneous data fusion for predicting mild cognitive impairment conversion81
A novel sparse regression method to fuse the auxiliary data into the predictor data for the pMCI/sMCI classification and shows the proposed method achieved the best classification performance, compared to the best comparison method, in terms of four evaluation metrics.
- Unsupervised meta-learning for few-shot learning80
An unsupervised meta-learning algorithm that learns from an unlabeled dataset and adapts to downstream human-specific tasks with few labeled data and constructs tasks using clustering embedding methods and data augmentation functions to satisfy two critical class distinction requirements is proposed.
- MWP-BERT: Numeracy-Augmented Pre-training for Math Word Problem Solving78
This work argues that injecting numerical properties into symbolic placeholders with contextualized representation learning schema can provide a way out of the dilemma in the number representation issue here and builds MWP-BERT, an effective contextual number representation PLM.
- 78
- 77
- Bounded coordinate system indexing for real-time video clip search75
This work proposes a novel video clip representation model called the Bounded Coordinate System (BCS), which is the first single representative capturing the dominating content and content—changing trends of a video clip and demonstrates that BCS achieves very high search accuracy according to human judgment.
- Teacher-Student Networks with Multiple Decoders for Solving Math Word Problem72
This paper proposes a novel approach, TSN-MD, by leveraging the teacher network to integrate the knowledge of equivalent solution expressions and then to regularize the learning behavior of the student network to addressMath word problem challenges.
- 72
- 70
- Is normalization indispensable for training deep neural network70
The proposed method can achieve the same or even slightly better performance in a variety of tasks: image classification in ImageNet, object detection and segmentation in MS-COCO, video classi-cation in Kinetics, and machine translation in WMT English-German, etc.
- Hyper-node Relational Graph Attention Network for Multi-modal Knowledge Graph Completion65
A novel network to incorporate different modal information with graph structure information for more precise representation of multi-modal knowledge graph, termed as hyper-node relational graph attention (HRGAT) network is proposed.
- A Federated Learning Approach for Privacy Protection in Context-Aware Recommender Systems62
This research proposed a federated learning-based privacy-preserving CF model for context-aware recommender systems that work with a user-defined collaboration protocol to ensure users’ privacy.
- 60
- Attention-based LSTM with Semantic Consistency for Videos Captioning60
This work proposes a novel end-to-end framework named aLSTMs, an attention-based LSTM model with semantic consistency, to transfer videos to natural sentences, and explores the correlation between multi-modal representations for generating sentences with rich semantic content.
- 59
- Small Object Detection Using Deep Feature Pyramid Networks58
A new two-stage detector similar to Faster-RCNN is developed, and dense convolutional network is used to strengthen the feature transmission and multiplexing, which leads to more accurate classification with fewer parameters.
- Efficient continuous top-k spatial keyword queries on road networks58
This paper investigates a novel problem, namely, continuous top-k spatial keyword queries on road networks, for the first time and proposes two methods that can monitor such moving queries in an incremental manner and reduce repetitive traversing of network edges for better performance.
- Content-Aware Summarization of Broadcast Sports Videos: An Audio–Visual Feature Extraction Approach57
This paper proposes a method for automatically generating highlights from broadcast sports videos by extracting audio and visual features from a sports video using SIFT features, and automatically learns the scorebox template from a broadcast sports video and locates and extracts thetemplate from a video stream.
- Traffic sign detection and recognition based on pyramidal convolutional networks57
This work develops a two-stage network that achieves a very great performance and surpasses the other state-of-the-art methods to implement a more comprehensive detection and recognition system for traffic signs.
- 56
- 56
- Personalized semantic trajectory privacy preservation through trajectory reconstruction53
An effective personalized trajectory privacy preserving method which can strike a good balance between user-defined privacy requirement and data availability in off-line trajectory publishing scenario and the results show that the method can improve the user trajectory availability as much as possible, while effectively achieving the different trajectory privacy requirements.
- Privacy-preserved and Responsible Recommenders: From Conventional Defense to Federated Learning and Blockchain51
The survey summarizes recent, high-impact works on diverse aspects and technologies that ensure responsible behavior in RS through an interconnected taxonomy and offers a holistic perspective on this rapidly evolving landscape by synthesizing insights from both RS and responsible AI literature.
- Lightweight dynamic conditional GAN with pyramid attention for text-to-image synthesis51
LD-CGAN is a compact and structured single-stream network that consists of one generator and two independent discriminators to regularize and generate 642 and 1282 images in one feed-forward process and significantly decreases the number of parameters and computation time.
- Breaking Isolation50
This work proposes EgoGCN, a structure that seeks to enhance multimodal learning of user-item interactions with a simple yet effective fusion operation dubbed EdGe-wise mOdulation (EGO) fusion that comfortably beats prior methods.
- Action-Centric Relation Transformer Network for Video Question Answering50
An action-centric relation transformer network (ACRTransformer) for VideoQA is proposed and a visual feature encoding technique, action-based encoding (ABE), is presented to emphasize the frames with high actionness probabilities (the probability that the frame has actions).
- E-FCNN for tiny facial expression recognition49
An edge-aware feedback convolutional neural network (E-FCNN) for tiny FER, which associates image super-resolution and facial expression recognition together and to effectively leverage the texture information of faces, is proposed.
- 48
- 47
- Learning What and When to Drop46
MetaDrop is presented, a differentiable and end-to-end approach for the ERC task that learns module-wise decisions across modalities and conversation flows simultaneously, which supports adaptive information sharing pattern and dynamic fusion paths.
- 46
- 43
- 43
- Destination prediction using deep echo state network42
Experimental results using the dataset provided by the Kaggle taxi trajectory prediction challenge show that the deepESN-DI and deepESn-DI-MLP approaches outperform other destination prediction models based on neural networks, namely MLP, RNN and the shallow ESN architectures.
- ECG Arrhythmia Classification using High Order Spectrum and 2D Graph Fourier Transform42
In this study, a feature extraction method based on the bispectrum and 2D graph Fourier transform (GFT) was developed, which achieved a high classification accuracy of 96.2%.
- Scalable Multimedia Retrieval by Deep Learning Hashing with Relative Similarity Learning41
A semi-supervised deep learning hashing (DLH) method for fast multimedia retrieval that utilizes both visual and label information to learn an relative similarity graph that can more precisely reflect the relationship among training data, and then generates the hash codes based on the graph.
- 41
- Zero-shot Cross-modal Retrieval by Assembling AutoEncoder and Generative Adversarial Network40
A novel approach named Assembling AutoEncoder and Generative Adversarial Network (AAEGAN) is proposed, which combines the strength of AutoEncoding and GAN to jointly incorporate common latent space learning, knowledge transfer, and feature synthesis for ZS-CMR.
- 40
- 40
- Clustering Web video search results based on integration of multiple features40
This paper presents the first system that clusters Web video search results by fusing the evidences from a variety of information sources besides the video content such as title, tags and description, and proposes a novel framework that can integrate multiple features and enable it to adopt existing clustering algorithms.
- 38
- Video-based person re-identification via spatio-temporal attentional and two-stream fusion convolutional networks38
A novel approach for video-based person re-identification using spatio-temporal attentional and two-stream fusion convolutional networks, which consist of the TSF-CNN, the long short-term memory networks, the spatial attention sub network and the temporal attention subnetwork is proposed.
- 37
- 37
- EEG-Based Emotion Recognition Using an Improved Weighted Horizontal Visibility Graph37
An attempt was made to extract complex network features from electroencephalogram (EEG) signals for emotion recognition using a novel method of constructing forward weighted horizontal visibility graphs (FWHVG) and backward weighted horizontal visible graphs (BWH VG) based on angle measurement.
- W-MAE: Pre-trained weather model with masked autoencoder for multi-variable weather forecasting36
Experimental results show that the W-MAE framework offers three key benefits: when predicting the future state of meteorological variables, the utilization of the pre-trained W-MAE can effectively alleviate the problem of cumulative errors in prediction, maintaining stable performance in the short-to-medium term and when predicting diagnostic variables, the model exhibits significant performance advantages over FourCastNet.
- Deep canonical correlation analysis with progressive and hypergraph learning for cross-modal retrieval36
A hypergraph semantic embedding method, which extracts latent semantics from texts, is introduced into DCCA to regularize the latent space learned by image view and text view, and a search-based similarity measure is proposed to score relevance of image-text pairs.
- 35
- Scenario-Aware Recurrent Transformer for Goal-Directed Video Captioning33
This work proposes a novel, simple but effective scenario-aware recurrent transformer (SART) model to execute video captioning, containing a “scenario understanding” module to obtain a global perspective across multiple frames, providing a specific scenario to guarantee a goal-directed description.
- Question-Led object attention for visual question answering32
This work proposes a novel but simple framework, namely Question-Led Object Attention (QLOB), to improve the VQA performance by exploring question semantics, fine-grained object information, and the relationship between those two modalities.
- Partial attention and multi-attribute learning for vehicle re-identification31
A partial attention and multi-attribute learning network to improve the discrimination and the robustness of vehicle re-identification and results show that the proposed method achieves higher accuracy compared with other methods.
- Image Tagging with Social Assistance30
A new image tagging scheme, termed social assisted media tagging (SAMT), which leverages the abundant user-generated images and the associated tags as the "social assistance" to learn the classifiers to solve the major challenges of noisy tags and desirable robustness of the tagging model.
- BlockDance: Reuse Structurally Similar Spatio-Temporal Features to Accelerate Diffusion Transformers29
This work proposes BlockDance, a training-free approach that explores feature similarities at adjacent time steps to accelerate DiTs, and introduces BlockDance-Ada, a lightweight decision-making network tailored for instance-specific acceleration.
- 29
- 29
- 29
- 28
- An efficient multi-scale channel attention network for person re-identification28
This work designed a novel cross-channel attention module (CCAM) in EMCA and placed it after different layers in the backbone and showed that the performance of the EMCA is consistently significantly superior to the existing state-of-the-art methods.
- Cognitive visual anomaly detection with constrained latent representations for industrial inspection robot28
A novel cognitive VAD method for industrial inspection robot is proposed based on a convolutional generative adversarial network and a denoising auto-encoder, where the adversarial learning mechanism is adopted to establish the boundary between the normal and anomaly data.
- Sharp and Real Image Super-Resolution Using Generative Adversarial Network28
A novel residual network architecture based on generative adversarial network (GAN) for image super-resolution (SR), which is capable of inferring photo-realistic images for 4\(\times \) upscaling factors, and demonstrates that the proposed approach performs better than previous methods.
- 27
- 27
- Kernel Attention Network for Single Image Super-Resolution27
This article proposes a novel kernel attention module (KAM) for SISR, which enables the network to adjust its receptive field size corresponding to various scales of input by dynamically selecting the appropriate kernel.
- 27
- 27
- 26
- Application of an Improved U2-Net Model in Ultrasound Median Neural Image Segmentation26
The data suggest the improved U2-Net model might be used for segmentation of ultrasound median neural images, and the ground truth came from the labeling of clinicians.
- TFNet: Multi-Semantic Feature Interaction for CTR Prediction26
A novel Tensor-based Feature interaction Network (TFNet) model is proposed, which introduces an operating tensor to elaborate feature interactions via multi-slice matrices in multiple semantic spaces and outperforms the competitive compared methods on the typical Criteo and Avazu datasets.
- Targeted Advertising in Public Transportation Systems with Quantitative Evaluation26
A probabilistic data model is proposed that captures the motion patterns and user interests so as to quantitatively evaluate the impact of an advertisement among the passengers and solves the retrieval problem efficiently to support real-time decision making.
- 26
- 25
- 25
- 25
- 25
- 24
- Augmented keyword search on spatial entity databases24
A new type of query is proposed that augments the spatial keyword search with an additional boolean expression constraint and an adaptive query processing strategy that takes into account the selectivity of query keywords and predicates for early termination is proposed.
- Personalized travel route recommendation from multi-source social media data24
The proposed approach combining multi-source social based topical package and long short-term memory (LSTM) based method is evaluated on a real travel dataset and compared with three state-of-the-art methods and shows that the method performs better for providing personalized travel routes.
- Statistical summarization of content features for fast near-duplicate video detection23
A system for detecting near-duplicate videos based on a novel summarization of content features for each clip that captures the dominating content and content changing trends of a video, so this representation is very compact and effective.
- 22
- Structure-driven protein engineering for production of valuable natural products22
Recent advances in protein engineering in synthetic biology, chemoenzymatic synthesis, and plant metabolic engineering are reviewed and opportunities for designing and constructing novel enzymes or proteins with desirable properties to obtain more active natural products are described.
- 22
- A coaxial couple wheeled robot with T‐S fuzzy equilibrium control22
This paper introduces a high load capacity coaxial couple wheeled robot (CCWR) and investigates a simple structure but effective fuzzy equilibrium controller based on (Takagi‐Sugeno) T‐S for balance control in wide‐angle range, which can achieve self‐balance and pivot around.
- Real-time image stabilization for arbitrary motion blurred image based on opto-electronic hybrid joint transform correlator22
The experimental results show that the proposed approach to keep real-time image stabilization based on opto-electronic hybrid processing has advantages of real time and preferable effect.
- 22
- 21
- Responsible Recommendation Services with Blockchain Empowered Asynchronous Federated Learning21
A responsible recommendation generation framework based on blockchain-empowered asynchronous FL that can be adopted for any model-based recommender system and evaluation results on three diversified benchmarks illustrate that the recommendation performance is considerably comparable with the standard and federated version of a well-known neural collaborative filtering recommender.
- 21
- PFFN21
A novel lightweight SR network named progressive feature fusion network (PFFN) is developed to seek for better balance between performance and running efficiency, with a novel progressive attention block (PAB) proposed as the main building block of PFFN.
- Video-based Person Re-identification via Self-Paced Learning and Deep Reinforcement Learning Framework21
A novel self-paced learning algorithm for video-based person re-identification is proposed, which could gradually learn from simple to complex samples for a mature and stable model and employ deep reinforcement learning to discard misleading and confounding frames and find the most representative frames from video pairs.
- 20
- Do Simpler Statistical Methods Perform Better in Multivariate Long Sequence Time-Series Forecasting?20
This work investigates two simple statistical methods for MLSTF and provides analysis to indicate that linear regression owns a lower upper bound of error than deep learning methods and SNaive can act as an effective nonparametric method with unpredictable trends.
- Finite-time stability of coupled impulsive neural networks with time-varying delays and saturating actuators20
Some sufficient conditions are obtained to ensure the stability of coupled impulsive neural networks with time-varying delays and saturating actuators in finite time by using polytopic representation approach and sector nonlinearity model approach.
- Adaptive Object Detection and Visibility Improvement in Foggy Image20
A novel algorithm to detect foreground objects from video sequences with fog and then enhance their visibilities and a novel metric to measure the image fog property to decide whether the image scene is obscured by fog or not is proposed.
- 20
- Mutual Correlation Network for few-shot learning19
A Mutual Correlation Network (MCNet) is proposed to explore global consensus among the correlation map by using the self-attention mechanism which has a global receptive field.
- Enhancing Radar Echo Extrapolation by ConvLSTM2D for Precipitation Nowcasting19
An enhanced convolutional long short-term 2D (ConvLSTM2D) based architecture for precipitation nowcasting is introduced, which includes time-distributed layers that enable parallel Conv2D operations on each image input, enabling effective analysis of spatial patterns.
- Extracting Useful Knowledge from Noisy Web Images via Data Purification for Fine-Grained Recognition19
This work investigates the noisy label problem and proposes a method that can specifically distinguish in- and out-of-distribution noisy samples and train the model on a less noisy web training set to achieve better robustness and performance.
- An Online Approach for Direction-Based Trajectory Compression with Error Bound Guarantee19
This paper investigates two approaches extended from Douglas-Peucker and Greedy Deviation algorithms respectively, and proposes a faster approximate algorithm with error bound guarantee named Angular-Deviation, which can achieve low running time to suit the most constrained computation environments.
- Examining User-Friendly and Open-Sourced Large GPT Models: A Survey on Language, Multimodal, and Scientific GPT Models18
This survey paper provides an examination of alternative open-sourced models of large GPTs, focusing on user-friendly and relatively small models that facilitate easier deployment and accessibility, and aims to equip researchers, practitioners, and enthusiasts with a thorough understanding of these models.
- 18
- Injection attack estimation of networked control systems subject to hidden DoS attack18
This paper is concerned with the injection attack estimation for a class of networked control systems with Deny-of-Service (DoS) attack, where unknown signals are injected into the sensor reading and actuator, and an estimator is designed to estimate the injected signals subject to stochastic hidden DoS attack.
- 18
- 18
- 18
- 18
- 18
- Factuality Matters: When Image Generation and Editing Meet Structured Visuals17
This work presents the first comprehensive, systematic investigation of structured visuals domain, encompassing data construction, model training, and an evaluation benchmark, and introduces StructBench, a novel benchmark for generation and editing with over 1,700 challenging instances.
- Communication-Efficient Federated Neural Collaborative Filtering with Multi-Armed Bandits17
This work proposes a communication-efficient neural collaborative filtering method for federated recommender systems that can achieve recommendation performance comparable with the standard federated neural collaborative filtering.
- 17
- 17
- Transductive Visual-Semantic Embedding for Zero-shot Learning17
This work proposes a novel framework termed transductive visual-semantic embedding (TVSE) for ZSL that achieves competitive performance compared with the state-of-the-arts for zero-shot recognition and retrieval tasks.
- 17
- Challenges and techniques for effective and efficient similarity search in large video databases17
An overview of the investigations on effective and efficient video similarity search is given and some novel techniques developed for two specific tasks studied in this PhD project: video retrieval in a large collection of segmented video clips, and video subsequence identification from a long unsegmented stream.
- Fully Functional Image Manipulation Using Scene Graphs in A Bounding-Box Free Way16
This method is designed to be the first to support fully functional image manipulation using scene graphs, including adding, removing, replacing and repositing instances, and significantly surpasses the state-of-the-art both quantitatively and qualitatively.
- Joint image deblurring and super-resolution with attention dual supervised network16
Extensive experiments indicate that the proposed attention dual supervised network (ADSN) not only generates remarkably clear HR images, but also achieves compelling results for joint image deblurring and SR task.
- MRMRP: Multi-source Review-Based Model for Rating Prediction16
A novel review-based model MRMRP, which stands for Multi-source Review-based Model for Rating Prediction, is proposed, capable of extracting useful features from supplementary reviews to further improve recommendation performance by applying a deep learning based method.
- An Efficient Online Approach for Direction-Preserving Trajectory Simplification with Interval Bounds16
- Unifying Multi-Source Social Media Data for Personalized Travel Route Planning16
The proposed multi-source topical package (MSTP) approach is evaluated on a real dataset and compared with two state-of-the-art methods, and the result shows that MSTP performs better for providing personalized travel routes.
- Locality condensation16
Locality Condensation (LC) is proposed, a novel idea to not only preserve localities determined by neighborhood information and their global similarity relationship, but also ensure that different localities will not invade each other in the low-dimensional subspace.
- 15
- 15
- 15
- 14
- LLMDiff: Diffusion Model Using Frozen LLM Transformers for Precipitation Nowcasting14
A probabilistic spatiotemporal model for precipitation nowcasting, named LLMDiff, which outperforms state-of-the-art models on the SEVIR dataset and enables the accurate estimation of motion trend by considering long-term temporal context information and capturing temporal dependencies within the frame sequence.
- 14
- 14
- Movie fill in the blank by joint learning from video and text with adaptive temporal attention14
This paper proposes to use adaptive temporal attention and fuse text information with attention to predict the word in the blank in a movie clip and description which has one blank.
- Towards Improving Canonical Correlation Analysis for Cross-modal Retrieval14
An improved CCA algorithm (ICCA) from three aspects is proposed, which proposes two effective semantic features based on text features to improve intra-modal semantic consistency and inspired by large scale similarity learning (LSSL), a similarity metric is proposed to improve the distance measure.
- Kernel based latent semantic sparse hashing for large-scale retrieval from heterogeneous data sources14
KLSSH firstly employs sparse coding for obtaining primary latent features of image and matrix factorization for generating features of text concepts to learn latent semantic features in a high level abstraction space and maps the latent semantic feature to compact binary codes using kernel method.
- Sensorless control for switched reluctance motor based on special position detection14
This paper proposes a new sensorless control approach for switched reluctance motor (SRM) in the single pulse control mode in high speed operation that uses the cross point position of transformer electromotive force (EMF) and motional back electromotiveforce (BemF) to estimate the rotor position.
- A graph based method for constructing popular routes with check-ins14
This paper presents a popular route construction method named GRID based on collective knowledge which mine the check-in records which contain the route’s attributes to divide the whole space into regions and find the POIs of each region.
- Image Thresholding with Cell-like P Systems14
A novel three-level thresholding approach for image segmentation based on celllike P systems is proposed in order to improve the computational efficiency of multilevel thresholding.
- Investigation of digital hologram watermarking with double binary phase encoding14
The affections of phase levels of binary phase holograms on the quality of extracting watermark from the watermarked image have been analyzed, and results show that the 2/2 phase level combination is the best scheme of watermarking than other multiple combinations.
- Training-Free Class Purification for Open-Vocabulary Semantic Segmentation13
FreeCP, a novel training-free class purification framework designed to address the challenges posed by class redundancy and visual-language ambiguity, significantly boosts segmentation performance when combined with other OVSS methods.
- 13
- 13
- 13
- 13
- Multi-view multi-label learning for image annotation13
A multi-view multi-label (abbreviated by MVML) learning algorithm, in which it makes full use of the complementarity among the views and the base learners of ensemble learning, leading to higher accuracy of image annotation.
- 13
- 12
- Unsupervised urban scene segmentation via domain adaptation12
An adversarial learning method for domain adaptation from the perspectives of both image-level and feature-level that adapts images from two domains to appear as if drawn from one domain and the latter attempts to learn similarities between the source and target domains in feature- level common space.
- TKGFrame: A Two-Phase Framework for Temporal-Aware Knowledge Graph Completion12
A novel two-phase framework called TKGFrame is proposed that achieves more accurate prediction results as compared to several state-of-the-art baselines and improves the performance of predicting missing temporal facts via solving a constrained optimization problem.
- Two-stage deep learning for supervised cross-modal retrieval12
The proposed novel two-stage deep learning method for supervised cross-modal retrieval is evaluated on two publicly available data sets, and the experimental results show that the method is competitive with state-of-the-art methods.
- 12
- Face identification with second-order pooling in single-layer networks12
A new face image representation method under the framework of single-layer networks, where the key component is the second-order pooling layer, which outperforms the state-of-the-art face identification performance by large margins.
- Multi-part sparse representation in random crowded scenes tracking12
A fully unsupervised tracking algorithm based on a multi-part local sparse appearance model based on the facts that only non-occluded segments of a target are effective in feature matching, while the occluded segments are the disturbed ones is proposed.
- Web Information Systems Engineering – WISE 2010 Workshops12
This book contains the carefully selected and reviewed papers presented at three satellite events that were held in conjunction with the 11th International Conference on Web Information Systems Engineering, WISE 2010, in Hong Kong, China, in December 2010.
- LitFM: A Retrieval Augmented Structure-aware Foundation Model For Citation Graphs11
LitFM is proposed, the first literature foundation model designed for a wide variety of practical downstream tasks on domain-specific literature, with a focus on citation information, and leverages a knowledge-infused LLM, fine-tuned through a well-developed instruction paradigm.
- 11
- 11
- 11
- Multi-Attention Audio-Visual Fusion Network for Audio Spatialization11
A novel multi-attention fusion network (MAFNet) based on the self-att attention mechanism to extract the spatial features related to the sound source in the video frames and fuse them into audio features well is designed.
- 11
- Stability property of impulsive inertial neural networks with unbounded time delay and saturating actuators11
Some sufficient conditions to ensure global asymptotic stability are obtained by using Lyapunov function with the matrix form of 2-norm to ensure the stability of impulsive inertial neural networks.
- Clustering-based multiple instance learning with multi-view feature11
This paper proposes a similarity-based method with clustering in a multi-view feature manner to solve MIL problems efficiently and demonstrates the effectiveness of the clustering-based MIL (CMIL) model.
- A Real-Time Algorithm for Small Group Detection in Medium Density Crowds11
A social force based collision avoidance model is built on each individual for goal direction prediction, and employ the predicted goal directions instead of traditional positions and velocities in collective motion detection to find group members in crowded scenarios.
- Pairwise Active Appearance Model and Its Application to Echocardiography Tracking11
The pairwise active appearance model (PAAM) is utilized in tracking the left ventricle contour in echocardiography and is utilized to obtain improved tracking results in terms of localization accuracy when compared with expert-specified contours.
- 10
- Dynamic Bi-layer Graph Learning for Context-aware Sequential Recommendation10
A novel Dynamic Bi-layer Graph learning (DBIG)-based framework for sequential recommendation, which learns the sequence dynamics of user behaviours and contexts and applies BERT to the graph embedding for understanding the user sequential behaviours in bidirectional ways and generating recommendations.
- 10
- Controllable image caption with an encoder-decoder optimization structure10
This work introduces a new feature extraction model, namely FVC R-CNN, to learn both the salient features and the visual commonsense features of controlable image caption, and proposes a novel MT-LSTM neural network for sentence generation.
- 10
- 10
- A System for Spatiotemporal Anomaly Localization in Surveillance Videos10
This demonstration proposes an anomaly detection algorithm called 2stream-VAE/GAN by embedding VAE and GAN in a two-stream architecture, which automatically locates anomaly based on a pre-trained model, which suits well for both streaming and local videos.
- 10
- 10
- Travel distance versus navigation complexity: a study on different spatial queries on road networks10
An evaluation to compare the effectiveness of easiest-to-reach neighbor query against a classic nearest neighbor query in a real-world setting is presented and how to achieve tradeoff between navigation complexity and travel distance for advanced navigational assistance is shown.
- 10
- 9
- StreamE: Learning to Update Representations for Temporal Knowledge Graphs in Streaming Scenarios9
Experimental results show that StreamE can achieve better performance than baseline methods with 100x faster in inference, 25X faster in training, and only 1/5 parameter size, which demonstrates its superiority.
- Inferring context with reliable collaborators: a novel similarity estimation method for recommender systems9
This research proposes a novel item-context similarity (ICS) model capable of adaptively generating reliable collaborators for a subject user on a subject item that is fused into a weighting model called contextually reliable collaborators (CRC) that considers the current item context.
- Deep learning based QoE evaluation for internet video9
A hybrid network model that integrates a deep neural network and an improved recurrent neural network for representation learning of view-level QoE evaluation is proposed and a graph-based ranking algorithm is proposed to conduct user behavior analyzing, which is useful to online decision of service providers.
- Exploiting score distribution for heterogenous feature fusion in image classification9
The effect of exploiting score distribution to address the feature fusion problem and proposing a novel method named score-distribution MKL (SD-MKL) for image classification, which significantly benefits the performance of image classification.
- Event Video Mashup: From Hundreds of Videos to Minutes of Skeleton9
A novel approach to automatically generate a unified short video from a collection of Web videos to describe the storyline of an event, using a submodular based content selection model that embodies both importance and diversity to depict the event from comprehensive aspects in an efficient way.
- Distribution-based similarity measures for multi-dimensional point set retrieval applications9
A novel collective gauge of relevance ranking is introduced by evaluating the probabilities that point sets are consistent with the same distribution of the query by evaluating two non-parametric hypothesis tests in statistics.
- 9
- 9
- 9
- 8
- 8
- DiffuLT: How to Make Diffusion Model Useful for Long-tail Recognition8
This paper proposes a new pipeline for long-tail (LT) recognition that utilizes the long-tailed dataset itself to generate a balanced proxy that can be optimized through cross-entropy (CE), and employs a randomly initialized diffusion model, trained exclusively on the long-tailed dataset, to synthesize new samples for underrepresented classes.
- 8
- 8
- 8
- TEVL: Trilinear Encoder for Video-language Representation Learning8
Experimental results demonstrate that TEVL can learn powerful video-text representation and achieve competitive performance on three downstream tasks, including multimodal video captioning, video Question Answering (QA), as well as video and language inference.
- 8
- 8
- Application of Axis Orbit Image Optimization in Fault Diagnosis for Rotor System8
The proposed method is reliable and effective for the identification of the axis orbit and aids in online monitoring and automatic identification of rotor system faults.
- 8
- Local Graph Embedding Based on Maximum Margin Criterion (LGE/MMC) for Face Recognition8
An efficient dimensional reduction and data classification method--local graph embedding method based on maximum margin criterion (LGE/MMC) for dimensional reduction, which is applied in face recognition, and the results of face recognition experiments on ORL, YALE and AR face databases demonstrate the effectivity of the proposed method.
- Real-time restoration of an image blurred by arbitrary motion using an opto-electronic hybrid joint transform correlator8
An efficient approach to restoring a blurred image in real time by using opto-electronic hybrid processing, by which an image motion vector can be effectively detected and an accurate point spread function was modeled rapidly.
- 8
- 7
- ENet: event based highlight generation network for broadcast sports videos7
A deep learning-based method capable of automatically generating highlights from a broadcast sports video based on important events and user preferences and the results of the experiments show the superiority of the proposed method.
- Video Representation Learning with Graph Contrastive Augmentation7
This work proposes a novel contrastive self-supervised video representation learning framework, termed Graph Contrastive Augmentation (GCA), by constructing a video temporal graph and devising a graph augmentation that is designed to enhance the correlation across frames of videos and developing a new view for exploring temporal structure in videos.
- Exploring Parameter Space with Structured Noise for Meta-Reinforcement Learning7
Experimental results demonstrate the superiority of ESNPS against a number of competitive baselines and utilizes meta-learning and directly uses meta-policy parameters, which contain prior knowledge, as structured noises to perturb the base model for effective exploration in new tasks.
- ViCO: A Training Strategy towards Semantic Aware Dynamic High-Resolution6
This work proposes Visual Consistency Learning (ViCO), a novel training algorithm that enables the model to represent images of varying semantic complexities using different numbers of vision tokens using different numbers of MLP connectors.
- ATLASS: An AnaTomicaLly-Aware Self-Supervised Learning Framework for Generalizable Retinal Disease Detection6
This work introduces an anatomically aware SSL framework that strategically integrates domain expertise through specialized masking of vital retinal structures during pretraining, enabling the development of clinically relevant feature representations without extensive labeled data.
- 6
- 6
- 6
- Parameter Matching of Energy Regeneration System for Parallel Hydraulic Hybrid Loader6
The results show that the PHHL with the optimization scheme can achieve 9.12% fuel saving, which is on the brink of the potential of brake energy recovery and verifies the feasibility of applying hydraulic hybrid technology on the loader.
- 6
- 6
- A resource-aware approach for authenticating privacy preserving GNN queries6
A resource-aware approach is proposed that supports effective location privacy preservation and efficient query result authentication with a less storage, communication and computation overhead and an authentication algorithm based on Voronoi diagram for spatial queries.
- 6
- Progressive framework for deep neural networks: from linear to non-linear6
A novel progressive framework is proposed to try to combine the stability of linear methods and the ability of learning complex and abstract internal representations of deep learning methods to provide a better and faster solution for deep neural networks.
- Dynamic User Attribute Discovery on Social Media6
This work presents a probabilistic method for dynamic user attribute discovery that automatically detects user attributes and models the dynamics using time windows and decay function, thereby facilitating more accurate recommendation.
- A Closed-loop Background Subtraction Approach for Multiple Models based Multiple Objects Tracking6
This paper discusses a closed-loop object detection and tracking method that uses the object tracking results to adjust the Gaussian Mixture Models’ parameters to extract foreground object pixels moreaccurately and the accuracy of both object Detection and tracking are improved withoutincreasing computational costs.
- Easiest-to-reach neighbor search6
This paper proposes an algorithm based on network expansion, which uses incremental processing of graph transformation that models instruction complexity and can efficiently find the easiest-to-reach neighbor with the guaranteed smallest navigation cost.
- 6
- Colony Algorithm for Wireless Sensor Networks Adaptive Data Aggregation Routing Schema6
A routing algorithm for data aggregation based on ant colony algorithm (ACAR) is presented, which is a distributed routing algorithm and realizes data aggregation trade-off in energy and delay.
- Cross-Modal Full-Mode Fine-Grained Alignment for Text-to-Image Person Retrieval5
FMFA, a cross-modal Full-Mode Fine-Grained Alignment framework, which enhances global matching through Explicit Fine-Grained Alignment (EFA) and existing implicit relational reasoning—hence the term “full-mode”—without introducing extra supervisory signals is proposed.
- Online Detection of Anomalies in Temporal Knowledge Graphs with Interpretability5
This work introduces AnoT, an efficient TKG summarization method tailored for interpretable online anomaly detection in TKGs, and demonstrates that AnoT surpasses existing methods significantly in terms of accuracy and interoperability.
- 5
- 5
- Strong-Weak Cross-View Interaction Network for Stereo Image Super-Resolution5
An efficient Transformer feature extraction block is introduced, which can efficiently capture long-range pixel interactions with lower resource consumption and the strong-weak cross-view interaction mechanism can be improved significantly with a negligible increment of computing cost.
- Laryngeal Diadochokinesis in the Mandarin Population5
According to the preliminary data, /ʔa/ and /ha/ can be used as suitable indicators to analyse the coordination ability of the laryngeal movement of Mandarin speakers.
- Understanding the Semantics of GPS-based Trajectories for Road Closure Detection5
This paper presents a novel road closure detection framework based on mining the semantics of trajectories, called T-Closure, which can detect 7000+ closure events monthly, with a delay within 1.5 hours.
- Generalized Unbiased Scene Graph Generation5
The Multi-Concept Learning (MCL) framework is proposed, which ensures a balanced learning process across rare/ uncommon/ common concepts and introduces the Balanced Prototypical Memory (BPM), which guides SGG models to generate balanced representations for concept-prototypes.
- Compositional Mathematical Encoding for Math Word Problems5
This work proposes a Compositional Math Word Prob-lem Solver (C-MWP) which works in a bi-modal setting encoding in an interactive way and shows its superiority over state-of-the-art models on public benchmarks.
- Pixel-level self-paced adversarial network with multiple attention in single image super-resolution5
This paper proposes a pixel-level self-paced adversarial network with multiple attention (PSPA) method to reduce the noise of SR image and increase its structural similarity with HR image through the combination of multiple attentions.
- Hierarchical pseudo-label learning for one-shot person re-identification5
This work proposes a novel approach targeting one-shot person re-identification that overcomes the information loss problem by introducing a hierarchical pseudo-label assignment strategy to fully exploit the unlabeled data information.
- 5
- An improved separation method of multi-components signal for sensing based on time-frequency representation5
An improved separation method based on the time-frequency distribution that can separate the signal whose components overlapped in TFR and have different time durations is proposed.
- Multi-view Semi-supervised Learning for Web Image Annotation5
This work proposes a novel model, termed as multi-view semi-supervised learning (MVSSL), for robust image annotation task, which exploits both labeled images and unlabeled images to uncover the intrinsic data structural information.
- 5
- Numerical Analysis of Modeling Based on Improved Elman Neural Network5
Simulation results of the half-bridge class-D power amplifier (CDPA) with two-tone signal and broadband signals as input have shown that the proposed behavioral modeling can reconstruct the system of CDPAs accurately and depict the memory effect ofCDPAs well.
- Max-margin adaptive model for complex video pattern recognition5
The proposed max-margin adaptive (MMA) model for complex video pattern recognition, which can utilize a large number of unlabeled videos to assist the model training, is proposed.
- 5
- 5
- Artificial intelligence method for automatic classification of vibration signals in the mining process4
An artificial intelligence-based automatic classification network model for mine vibration signals based on conventional convolutional neural networks and which incorporates long short-term memory (LSTM) networks and attention mechanisms is proposed.
- 4
- 4
- 4
- 4
- RICAPS4
Residual inception cascaded Capsule Network (RICAPS) significantly improves the performance of broadcast sports video classification as deeper features are captured by the cascaded capsule network.
- Multiplicative angular margin loss for text-based person search4
Text-based person search aims at retrieving the most relevant pedestrian images from database in response to a query in form of natural language description by proposing multiplicative angular margin (MAM) loss to learn angularly discriminative features for each identity.
- Context Encoding for Video Retrieval with Contrastive Learning4
This paper proposes CECL (Context Encoding for video retrieval with Contrastive Learning), a video representation learning framework that aggregates the context information of frame-level descriptors, and a supervised contrastive learning method that performs automatic hard negative mining, and utilizes the memory bank mechanism to increase the capacity of negative samples.
- 4
- 4
- Movie Fill in the Blank with Adaptive Temporal Attention and Description Update4
This paper proposes to use a novel LSTM network called L STM with Linguistic gate (LSTMwL), which exploits adaptive temporal attention for MovieFIB, and demonstrates that this approach outperforms state-of-the-art models for movieFIB.
- 4
- Exploring Viewable Angle Information in Georeferenced Video Search4
This study proposes a novel circular sector model with beginning-ending vectors for FOV representation which additionally explores viewable angle information and demonstrates how the model can be applied to perform different types of overlap queries for spatial data selection in a unified framework, while providing competitive performance in terms of efficiency.
- Geographical Constraint and Temporal Similarity Modeling for Point-of-Interest Recommendation4
Improved collaborative filtering models are introduced by jointly taking advantages of geographical constraint and temporal similarity to capture personalized geographical constraint of each individual.
- 4
- An Improved Model Based on BPNN for Class-D Power Amplifier4
An improved behavioral model based on BP neural network and approximation theory is proposed to describe the nonlinearity of Class-D power amplifier, which has faster computing speed and higher accuracy than traditional model.
- 4
- The Implementation of High-speed Floating-point DBF Based on FPGA4
Comparison of the proposed approach with another approach based on the Xilinx IP core generator modules shows that the former technique outperforms the latter in terms of latency and speed.
- 4
- 4
- Estimation of contour motion and deformation for nonrigid object tracking4
This work solves the nonrigid contour tracking problem by decomposing it into three subproblems: motion estimation, deformation estimation, and shape regulation, and uses a shape prior model to constrain the deformed contour.
- 4
- A Responsible and Extendable Context-Aware Recommender System3
This work designs RECARS, the first responsible and extendable context-aware recommender system that allows users to perform iteratively refining and explaining the results by active learning and large language model (LLM).
- Comparative Analysis and Suggestions for Sucker Rod Pump Working Condition Diagnosis Using Machine Learning Techniques3
A significant inverse correlation between model accuracy and the number of trainable parameters is demonstrated, notably with EfficientViT achieving the highest accuracy despite having fewer parameters, suggesting that simpler, but more efficient models can outperform more complex architectures in specific tasks like SRP diagnosis.
- Multi-scale Consistency Deep Lifelong Cross-modal Hashing3
A regularization-based lifelong learning module is introduced to balance plasticity for learning new knowledge and stability for maintaining old knowledge and update incremental hash codes without retraining cumulative data, and multi-scale consistency network which employs multi-scale feature fusion module is introduced to learn multi-level semantic representations with consistency.
- 3
- ESTISR: Adapting Efficient Scene Text Image Super-resolution for Real-Scenes3
This work designed a re-parameterized inverted residual block specifically suited for resource-limited circumstances as the feature extractor and proposed a novel self-attention mechanism, softmax shrinking, based on a kernel-based approach, which consistently outperforms current methods in terms of actual running time and peak memory consumption.
- 3
- 3
- 3
- 3
- 3
- 3
- Document Binarization using Recurrent Attention Generative Model.3
Evaluated on various standard benchmarks, the proposed method outperforms state-of-the-art binarization methods both quantitatively and qualitatively and performs well in the task of image unshadowing, which evidently verifies its generality.
- Multi-class Human Body Parsing with Edge-Enhancement Network3
An edge enhancement module to emphasize the potential of edge loss and boundary information is proposed and local and global information will be explored for complex multi-class human body parsing problem by densely connected atrous spatial pyramid pooling.
- 3
- 3
- Accelerating spatial join operations using bit-indices3
An innovative method for accelerating spatial join operations using Spatial Join Bitmap (SJB) indices is proposed and results show that the use of SJB indices produces substantial speedup ranging from 25% to 150% when compared to Filter trees.
- 3
- 3
- Recursive partitioning method for trajectory indexing3
This paper provides the methodology for using the recursive partitioning technique for indexing trajectories in the unrestricted space, which is called the Recursively Partitioned Trajectory Index (RPTI).
- Dual Dimensionality Reduction for Efficient Video Similarity Search3
This paper proposes a novel video reduction method called Optimal Dual Dimensionality Reduction (ODDR) to dramatically reduce the video data complexity for accurate and quick search, by reducing the dimensionality of both feature vector and sequence.
- Towards Pattern-Aware Data Augmentation for Temporal Knowledge Graph Completion2
A hierarchical scoring algorithm based on triadic closures within TKGs is proposed, which enables pattern-aware validation for new samples and proposes a two-stage training approach to identify samples that deviate from the model's preferred patterns.
- 2
- Online Anomaly Detection over Live Social Video Streaming2
A novel deep neural network model called Coupling Long Short-Term Memory (CLSTM) that adaptively captures the history behaviours of the presenters and audience, and their mutual interactions to predict their behaviour at next time point over streams is proposed.
- 2
- 2
- Abstract and Explore: A Novel Behavioral Metric with Cyclic Dynamics in Reinforcement Learning2
A novel Behavioral metric with Cyclic Dynamics (BCD) is proposed, which considers both cumulative and immediate effects and facilitates the abstraction and exploration of the agent and provides a thorough awareness of environmental dynamics.
- 2
- 2
- 2
- 2
- 2
- 2
- 2
- 2
- Web and Big Data2
The results on real-world dataset show the framework can apply to large-scale datasets, help the user conduct subsequent data mining work, and help the user cognize the dark data.
- Dual-Path Recurrent Network for Image Super-Resolution2
This work proposes a novel architecture based on dual-path block (DPB) that enhances the network performance simply by increasing the number of DPB recursions and avoiding additional parameters, and exploits the relationship between the two states by introducing a back-projection operation.
- High Order Neural Networks for Video Classification.2
This paper presents high order operations as a generic family of building blocks for capturing high order correlations from high dimensional input video space, and proposes a new LEarnable hiGh Order (LEGO) block, whose goal is to capture spatiotemporal correlation in a feedforward manner.
- 2
- Public Security Video and Image Analysis Challenge: A Retrospective2
A review of the Public Security Video and Image Analysis Challenge, including tasks definition, datasets creation, ground truth annotation, and results comparison and analysis is provided.
- 2
- Retrieving Images by Multiple Samples via Fusing Deep Features2
A novel image retrieval paradigm that users could input two images as query to search the images that include the content of the two input images-synchronously by proposing the FWC Feature weighting by Clustering, a novel algorithm to weight the two query features.
- Spatial Verification via Compact Words for Mobile Instance Search2
This paper proposes a novel approach for mobile instance search, by spatial analysis with a few representative visual words extracted from multi-photos, and demonstrates a notable improvement on average precision over several state-of-the-art methods including spatial coding, query expansion and multiple photos.
- 2
- SELSH: A Hashing Scheme for Approximate Similarity Search with Early Stop Condition2
This paper proposes a new scheme called SortedKey and Early stop LSH (SELSH), which extends the previous SortingKeys-LSH (SK- LSH), which uses a linear order to sort all the compound hash keys.
- 2
- 2
- 2
- 1
- 1
- 1
- 1
- 1
- Large Language Models' Complicit Responses to Illicit Instructions across Socio-Legal Contexts1
The authors' findings reveal widespread LLM susceptibility to complicit facilitation, with GPT-4o providing illicit assistance in nearly half of tested cases and existing safety alignment strategies are insufficient and may even exacerbate complicit behavior.
- 1
- 1
- 1
- 1
- 1
- 1
- 1
- 1
- 1
- 1
- 1
- 1
- 1
- Cross-modality consistency learning for visible-infrared person re-identification1
ACross-modality consistency learning network is proposed, which jointly considers cross-modal learning and distillation learning and could adaptively select the high-quality person image according to the potential contribution of each image, to avoid negative knowledge transfer.
- Action Anticipation in First-Person Videos with Self-Attention Based Multi-Modal Network1
A self-attention based multi-modal LSTM framework for the challenging task of action anticipation in first-person videos that outperforms the state-of-the-art approaches on metrics, and solved the problem of poor long-term prediction.
- Design of low-power CIC decimation filter based on nibble serial arithmetic1
A folded multistage cascade CIC decimation filter has a programmable structure of down sampling factor and can dynamically realize 1∼16 times extraction and the designed truncation module can optimize the effective accuracy of the final stage output.
- Cross-time transportation network improvement strategies under supply and demand uncertainty1
This paper considers the network optimization problem under the condition of cross-time decision-making and gives the network design model of stochastic bi-level programming, which is a genetic algorithm with Monte Carlo used to solve the cross- time transportation network design problem under supply and demand uncertainty.
- 1
- 1
- 1
- 1
- 1
- Guest editorial: special issue on Web and big data 20191
This special issue of World Wide Web on Web and big data 2019 presents five extended papers, including a novel method for matching addresses which has important applications in location-based businesses, by getting rich contexts through search engines, and a system to aid users in quickly discerning salient connections and factsWorld Wide Web.
- Adaptive Feature Enhancement Network for Semantic Segmentation1
An Adaptive Feature Enhancement Module (AFEM) that utilizes high-level features to adaptively enhance the key areas of low- level features and an Adaptive Features Enhancement Network (AFENet) is designed with AFEM to combine all the enhanced features.
- Towards Good Practices for Video Object Segmentation1
This work conducts a series of refinements with the propagation-based video object segmentation method and empirically evaluates their impact on the final model performance through ablation study to improve the space-time memory networks.
- A Method of Multi-component Signal Detection Based on Differential Nonlinear Mode Decomposition1
Proposed method can detect multi-component signals of time-varying amplitude without fake frequency under the condition of noise and chaos and the peaks of Hilbert marginal spectrum of proposed method are sharper, and R2, R3 are higher.
- Feature Reconstruction by Laplacian Eigenmaps for Efficient Instance Search1
This work focuses on reconstructing a new representation based on conventional CNN features to capture the intrinsic image manifold in the original feature space to capture implicit embedding information in images as well as reducing the computational complexity significantly.
- 1
- A Vehicle Model Data Classification Algorithm Based on Hierarchy Clustering1
A vehicle model data classification algorithm based on hierarchy clustering is proposed for the development of vehicle brand, style and year recognition products whose car face or tail has the same structure which can’t be distinguish in image or video.
- A Robust Image Super-solution Algorithm Using Deep Convolution Networks1
This work explores the architecture of convolutional neural network (CNN) and puts forward a robust SISR algorithm based on CNN, which obtains the best SR performance.
- 1
- Incomplete Surface Reconstruction for Statue Point Cloud based on Symmetry1
A novel method of reconstruction from incomplete statue point cloud that complete the missing parts of statue models before reconstruction using the symmetry of statue model, and the point cloud is reconstructed by Power Crust algorithm.
- 1
- An Improved Modeling of Nonlinear Circuits based on Elman Neural Network1
An improved modeling based on Elman neural network was proposed to analyze the nonlinear features of nonlinear circuits with the memory effect of PAs and could be also applied to analyze linear circuits and RF power amplifiers.
- Modeling based on Elman Wavelet Neural Network for Class-D Power Amplifiers1
The Foundation of Key Laboratory of China’s Education Ministry and A Project Funded by the Priority Academic Program Development of Jiangsu Higher Education Institutions are revealed.
- 1
- Web Information Systems Engineering - WISE 2010 Workshops: WISE 2010 International Symposium WISS, and International Workshops CISE, MBC, Hong Kong, ... Applications, incl. Internet/Web, and HCI)1
This book contains the carefully selected and reviewed papers presented at three satellite events that were held in conjunction with the 11th International Conference on Web Information Systems Engineering, WISE 2010, in Hong Kong, China, in December 2010.
- Design of High-speed Floating-point Arithmetic on FPGAs1
An improved dual-path architecture is used to reduce the latency of critical path and the logical function of floating point multiplier is simplified and the throughput of the system is increased.
- 1
- Simulation of the Asynchronous Motor System Based on Matlab/Simulink1
This paper presents the Matlab/Simulink-based mathematical model of the rotor flux orientation vector control system of an AC Asynchronous motor and the modules of the hysteresis current PWM controller and of the helicopter flux linkage observer.
- 1
- 1
- 1
- Heli-Tele: Road Extraction from Helicopter Video1
A learning based road likelihood computation method that uses aerial imagery and fuses information from several weak features to align a color helicopter image onto a given street map to find address of buildings visible in the image.
- IMAGE COMPRESSION FOR STORAGE TRANSMISSION OF DIGITAL IMAGES1
MrSID, a commercial software package for image compression, was chosen as a suitable method for the needs of MnDOT's Office of Land Management because of its main practical advantages, which include improved utilization of storage and transmission resources and a multi-resolution browsing capability.
- Temporal Dependency‐Aware Trajectory‐Level Behavioural Metric for Exploration in Reinforcement Learning–
This work proposes a novel novel trajectory representation framework that exploits temporal dependencies of the trajectory and captures the underlying sequential information of behaviour patterns and develops an auxiliary representation regulariser to promote the diversity and informativeness of the trajectory representation.
- –
- GPU and CPU Memory Co-Optimization in Heterogeneous Pipeline Parallelism for Efficient Large Language Model Fine-Tuning on Commodity Servers–
Tiny-Pipe comprises a holistic layer packing method that simultaneously reduces GPU memory footprint and improves training performance, an active CPU memory management that alleviates CPU memory pressure by eliminating redundant parameters, and a layer-wise runtime swapping strategy that further enhances overall performance.
- Enhancing NWP precipitation forecasting with Rain-PostEFNet: A multi-scale deep learning approach–
Rain-PostEFNet is proposed, a multi-scale deep learning framework for post-processing NWP precipitation forecasts that captures meteorological features across scales and mitigates severe class imbalance in extreme precipitation forecasts.
- –
- –
- –
- Task augmentation via channel mixture for few-task meta-learning–
This work introduces a novel task augmentation method called Task Augmentation via Channel Mixture (TACM), which generates new tasks by mixing channels from different tasks and preserves the continuity and integrity of feature information in channels during the mixture process, thereby enhancing the generalization ability of the model.
- Rethinking the Denoising Strategy in Session-Based Recommendation via Bidirectional Information Flow–
The Hybrid Prototype-based In-and-Out Flow Network (HyPro), which employs both denoising and enhancing processes for each session based on the authors' proposed hybrid prototypes, and demonstrates that HyPro outperforms the state-of-the-art baselines.
- Motivation-Aware Session Planning over Heterogeneous Social Platforms–
This paper proposes a Motivation-Aware Session Planning (MASP) framework for session planning over heterogeneous social platforms with a novel HeterBERT model to handle the heterogeneity of items at both type and attribute levels, and proposes to predict user preference using the motivations behind user activities.
- –
- –
- –
- –
- MemRoadNet: Human-like Memory Integration for Free Road Space Detection–
MemRoadNet is introduced, a Memory-Augmented semantic segmentation framework that integrates human-inspired cognitive architectures with deep-learning models for free road space detection and achieves competitive performance with complex multimodal systems while maintaining computational efficiency and achieving top performance among single-modality methods.
- –
- Style-unaware meta-learning for generalizable person re-identification–
A style-unaware meta-learning method that is less affected by domain migration is proposed, and a style fusion plugin is designed that changes the specific source domain’s style and simulates more diverse domain differences.
- –
- Diffusion Product Quantization–
This work proposes a method to compress the codebook by evaluating the importance of each vector and removing redundancy, ensuring the model size remaining within the desired range, and applies this method to the DiT model on ImageNet, demonstrating competitive generative performance.
- Automatic design system for digital up-conversion and down-conversion–
An automatic platform for automatically generating DUC/DDC RTL code that meets the design requirements through simulation and verification, which greatly improves the design efficiency of DUC and DDC.
- –
- –
- –
- –
- Medical image denoising based on 2D adaptive compact variational mode decomposition–
This paper proposes a medical image denoising method based on two-dimensional adaptive compact variational mode decomposition (2D-ACVMD), which can effectively remove the noise interference, retain the original edge and detail features, and the PSNR and SSIM obtained are higher than the previous methods.
- –
- Research on Surge Suppression of Timing Fuze Timing Starting Point Based on NMOSFET Device–
The value of the parallel capacitor CGD’ of the designed transient surge suppression circuit is studied, which can effectively suppress the surge, thus ensuring the accuracy of the timing starting point of the time fuze.
- –
- –
- –
- Controllable Image Caption Based on Adaptive Weight and Optimization Strategy–
An encoder-decoder structure is adopted, in which the encoder is used to detect the regions of interest of the image while the decoder is applied to generate sentences, and an adaptive weight learning algorithm is proposed, which can make the multiple loss functions better participate in training.
- Stochastic multi-objective model in traffic network design–
A stochastic multi-objective network design model with the concept of Pareto optimal is established that can provide scientific, systematic and effective decision-making scheme and decision data for the decision-maker.
- Discrete traffic network design model considering risk attitude of decision maker and users–
The results of the Nguyen Dupuis network show that the risk attitudes of decision makers and users have an important impact on the decision-making of network design, and the proposed method can effectively improve the robustness of networkDesign.
- –
- –
- –
- Learnable Higher-Order Representation for Action Recognition–
This paper presents learnable higher-order operations as a generic family of building blocks for capturing spatiotemporal dynamics from RGB input video space and shows that on the task of video recognition, these higher-order models can achieve results on par with or better than the existing state-of-the-art methods.
- –
- –
- –
- An Enhanced Method of Trajectory Privacy Preservation Through Trajectory Reconstruction–
An enhanced trajectory privacy preserving method is proposed which can protect the trajectory privacy preferably while maintaining a high utility of the trajectory in data publishing and an effective trajectory reconstruction algorithm is introduced to avoid the mutations of trajectory.
- –
- A Multi-Robot Path Planning Method under Narrow Environments–
This paper presents a dynamic allocation priority method based on learning classifier to improve the performance of the robot team and improves the capacity of multi-robot path planning.
- Robot reinforcement learning accuracy-based learning classifier systems with Fuzzy Policy Gradient descent(XCS-FPGRL)–
A novel approach XCS-FPGRL is presented which combines covering operator and genetic algorithm which acts as an innovation discovery component which is responsible for discovering new better reinforcement learning rules.
- –
- –
- –
- Multi-Robot Reinforcement Learning Based on LCS and LS-SVM *–
Algorithm analysis and simulation show that a large space for multi-robot learning, the learning speed converges slowly, uncertainties and other learning problems can get a great improvement.
- –
- A Hierarchical Conflict Resolution Method for Multi-Robot Path Planning–
This paper presents a learning classifier based on dynamic allocation of priority methods to improve the performance of the robot team and improve the capacity of multi-robot path planning.
- –
- Robot Reinforcement Learning Methods Based on XCSG–
This paper proposed a robot reinforcement learning method based on learning classifier system, a accuracy-based machine learning system with gradient descent that combines reinforcement learning and rule discovery system.
- Research on Multi-robot Path Planning Methods Based on Learning Classifier System with Gradient Descent Methods–
This paper deals with the problem of multi-robot path planning based on learning classifier system in a dynamic narrow environment, where the workspace is cluttered with unpredictably moving objects.
- Context Sensitive Tag Expansion with Information Inference–
The proposed approach uses modified Hyperspace Analogue to Language (HAL) model instead of association rules or latent dirichlet allocations to mine the correlations between image tags to overcome the limitation caused by the tag co-occurrence based methods.
- Method of Image Fusion Based on Improved Probability Theory–
This essay brings forward to a skill of compromising use ratio and image which can efficiently decrease interference made in compromising pictures and provides more abundant, accurate and reliable information according to physical characteristics generated by sensors, so further analysis and management can be done.
- A Multi-view Fuzzy Matching Strategy with Multi-planar Homography–
A fuzzy matching strategy using a multi-planar homography constraint that is ideally suitable for conditions with not enough features, and able to accurately match target blobs in all views.
- Robot Hybrid Architecture Based on Parallel Processing of Multi-DSP–
A new robot hybrid architecture based on that multi-DSP system, which combines different system structure adopted in different periods for development of robot, and which enhanced the entire robot assembly system’s Practicality, Real-time and Reliability is proposed.
- UQMSG experiments for TRECVID 2011–
Evaluation results suggest that more efforts need to be made to well tune system parameters, and sophisticated techniques beyond applying keyframe-level semantic concept propagation and near-duplicate detection are required for achieving better performance in video tasks.
- –
- Object Tracking and Mensuration in Surveillance Videos–
The first part of the thesis discusses several object tracking approaches based on the different properties of tracking targets and proposes a heightmeasuring algorithm with less human supervision, more flexibility and improved robustness.
- Robot Reinforcement Learning Based on Learning Classifier System–
This paper proposed a robot reinforcement learning method based on learning classifier system, which can easily reduce learning space, improve online learning ability and robustness.
- –
- Lecture Notes in Computer Science: Research on Multi-robot Avoidance Collision Planning Based on XCS–
A novel approach to solving the problem of robot path planning that combines covering operator and genetic algorithm and its accuracy-based representation that can easily reduce learning space, improve online learning ability, robustness due to the use of genetic algorithm.
- –
- A Data Processing Method Based on Cubic Spline Surface in Multibeam Systems–
The following simulations and analysis proved that this modified model-cubic spline interpolation was an effective method to modify the measurement of multibeam data, and widened the thoughts about outliers processing.
- –
- Multi-robot Reinforcement Learning Strategy Based on ILCS–
The algorithm analysis and the simulation indicate that the improved learning classifier system used in the multi-robot reinforcement learning is feasible and effective.
- Effective Construction of Learning Community in Classroom——"High-level Language Programming (VB)" Course as an Example–
Through the practice in high-level language programming course for two years, it is shown that Learning Community not only guaranteed the collaboration and communication in a limited classroom period to promote the students’ ability of absorbing knowledge, but also resolved the entry difficulty in learn programming course effectively.
- Effective and Efficient Similarity Search in Video Databases–
A Dynamic Query Ordering (DQO) algorithm for efficiently processing Batch Nearest Neighbor (BNN) search in high-dimensional space, with advanced optimizations of both I/O cost and CPU cost is proposed.
- –
- FIM Beamforming for Sparse Array–
In this paper interpolation technique is used for processing the FIM (Fourier Integral Method) beamforming of equispaced linear sparse array and improves the performance of directivity and noise rejection.
- Research Advance on Biological Pesticides–
After reviewing the present researching situation, biological pesticides was thought to have a bright prospect and the developing direction and major project of biological pesticides in China were suggested.
- Research on the Algorithm of Opportune Firing for Air Defense Firepower Unit–
According to the requirement of firepower assignment of modern antiaircraft, it reconstruct the index system that ascertain the air defense firepower unit opportune firing, and introduce the process of opportsune firing algorithm in the base of certain firing availability coefficient, firing availability times and killing probability.
- –
- –
- Laser direct writing system of fabricating high resolution image with DMD–
A method of using DMD(igital micro device) to carry out the high resolution image with high speed to engrave is put forward and experimental results have been given.
- The University of Queensland at TRECVID 2007 Search Task–
It is observed that there is no signifficant difference between these two runs in terms of what measured, so more sophisticated approaches such as multi-modality fusion are needed to improve the system performance.
- –
- –
- –
- Chemistry Experiment and Development of Students' Quality–
Through the exploration of chemistry experiments and development of students' qualities, the methods and requests in teaching practice so as to fully improve students' quality are outlined.
- A Real Time Adaptive Method and Its Application to the Azimuth Preprocessing of SAR Signal–
An adaptive azimuth preprocessing method is put forwand, which can greatly overcome theAzimuth modulation of SAR signal and can a meliorate the SNR of the SAR system, the auto-focus performance will be improved greatly.
- –
- –
- Intelligence generator of arbitrary wave–
This article gives an introduction to the way of making a high-quality and multi-founction signal generator by using a 8098 microcontroller as well as to the hardware structure of the generator.
- –
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.