Is this you? Claim this profile to correct it, add a bio and choose the work people see first.
Claim this profileWorks350 from public data
- Sequential Click Prediction for Sponsored Search with Recurrent Neural Networks355
A novel framework based on Recurrent Neural Networks (RNN) is introduced that directly models the dependency on user's sequential behaviors into the click prediction process through the recurrent structure in RNN.
- 281
- 273
- Knowledge Graph Embedding With Iterative Guidance From Soft Rules247
Experimental results show that with rule knowledge injected iteratively, RUGE achieves significant and consistent improvements over state-of-the-art baselines; and despite their uncertainties, automatically extracted soft rules are highly beneficial to KG embedding, even those with moderate confidence levels.
- Query dependent pseudo-relevance feedback based on wikipedia245
This work proposes and proposes and studies the effectiveness of three methods for expansion term selection, each modeling the Wikipedia based pseudo-relevance information from a different perspective, and incorporates the expansion terms into the original query and uses language modeling IR to evaluate these methods.
- Real-time Multi-Class Helmet Violation Detection Using Few-Shot Data Sampling Technique and YOLOv8216
This study proposes a robust real-time helmet violation detection system that utilizes a unique data processing strategy, referred to as few-shot data sampling, to develop a robust model with fewer annotations, and a single-stage object detection model, YOLOv8 (You Only Look Once Version 8), for detecting helmet violations in real- time from video frames.
- A probabilistic model for retrospective news event detection205
This work proposes a probabilistic model to incorporate both content and time information in a unified framework for retrospective news event detection and builds an interactive RED system, HISCOVERY, which provides additional functions to present events, Photo Story and Chronicle.
- Exploiting parallel texts for word sense disambiguation182
This paper evaluates an approach to automatically acquire sense-tagged training data from English-Chinese parallel corpora, which are then used for disambiguating the nouns in the SENSEVAL-2 English lexical sample task.
- Improving Knowledge Graph Embedding Using Simple Constraints158
The potential of using very simple constraints to improve KG embedding is investigated, examining non-negativity constraints on entity representations and approximate entailment constraints on relation representations to improve model interpretability.
- Question-answer topic model for question retrieval in community question answering127
A novel Question-Answer Topic Model (QATM) is proposed to learn the latent topics aligned across the question-answer pairs to alleviate the lexical gap problem, with the assumption that a question and its paired answer share the same topic distribution.
- Relation Extraction with Multi-instance Multi-label Convolutional Neural Networks.125
This paper proposes a multi-instance multi-label convolutional neural network for distantly supervised RE, which first relaxes the expressed-at-least-once assumption, and employs cross-sentence max-pooling so as to enable information sharing across different sentences.
- A novel refinement approach for text categorization113
The results indicate that in each case, the refined classifiers achieve significant performance improvement over the base classifiers used, and the performance of the Refined Centroid Classifier implemented is comparable, if not better, to that of state-of-the-art support vector machine (SVM)-based classifier, but offers a much lower computational cost.
- 95
- Constructing and Embedding Abstract Event Causality Networks from Text Snippets93
Experimental results demonstrate that the abstract causality network is effective for discovering high-level causality rules behind specific causal events, and the embedding models perform better than state-of-the-art link prediction techniques in predicting events.
- Simulation and control of skeleton-driven soft body characters91
A physics-based framework for simulation and control of human-like skeleton-driven soft body characters that couple the skeleton dynamics and the soft body dynamics to enable two-way interactions between the skeleton, the skin geometry, and the environment.
- Global Dynamic Path Planning Fusion Algorithm Combining Jump-A* Algorithm and Dynamic Window Approach82
The experimental results show that the new fusion algorithm proposed in this paper can not only effectively solve the problem of non-continuous curvature and excessive turning angle at the turning points of the path planned by jump-A* algorithm, but also improve the smoothness of the course and the global optimality.
- 68
- 66
- Automatic Classification of Cervical Cells Using Deep Learning Method65
The potential of a deep learning model to automatically distinguish abnormal cells from normal cells is explored in cervical cancer screening data collected from the fourth central hospital of Baoding city, China.
- Learning to rank for question routing in community question answering63
A general framework based on the learning to rank concepts for QR is proposed, which can both outperform the traditional query likelihood language model (QLLM) as well as the state-of-the-art Latent Dirichlet Allocation based model (LDA).
- 61
- 60
- Dual role model for question recommendation in community question answering57
This paper is the first to systematically investigate the distinctions between the two roles and their different influences on the performance of question recommendation in CQA, and proposes a Dual Role Model (DRM) to model the dual roles of users effectively.
- 53
- 50
- 50
- 48
- A study on mutual information-based feature selectionfor text categorization48
It is shown that the corrected MI method’s performance is similar to that of IG, and it is considerably better than PMI, and an MI method derived correctly from information theory is detail.
- 46
- 43
- Improved latent concept expansion using hierarchical markov random fields40
A novel query expansion technique that models the various types of dependencies that exist between original query terms and expansion terms within a robust, unified framework called Hierarchical Markov random fields (HMRFs), based on Latent Concept Expansion (LCE).
- 40
- 39
- 38
- View-based 3D object retrieval with discriminative views37
This paper focuses on the 2D projective views that represent the 3D objects and proposes a boosting approach by evaluating the discriminative ability of each objects views to boost the multi-model graph learning based retrieval method.
- EI-MTD: Moving Target Defense for Edge Intelligence against Adversarial Attacks36
Results show that EI-MTD could protect edge intelligence effectively against adversarial attacks in black-box settings.
- 34
- 34
- Text Adversarial Attacks and Defenses: Issues, Taxonomy, and Perspectives32
This work introduces the pipeline of NLP, including the vector representations of text, DNN-based victim models, and a formal definition of adversarial attacks, which makes the review self-contained.
- GazeSAM: What You See is What You Segment31
This study presents the \textbf{GazeSAM} system to enable radiologists to collect segmentation masks by simply looking at the region of interest during image diagnosis, and utilizes the eye-gaze data as the input prompt for SAM, which automatically generates the segmentation mask in real time.
- Domain Generalization with Correlated Style Uncertainty30
This work introduces a novel augmentation approach, named Correlated Style Uncertainty (CSU), surpassing the limitations of linear interpolation in style statistic space and simultaneously preserving vital correlation information.
- A Robust Wireless Sensor Network Localization Algorithm in Mixed LOS/NLOS Scenario30
In GMDS, the RSSI is estimated using a Gaussian mixed model based non-metric Multidimensional (GMM), the dissimilarity matrix is built to generate relative coordinates of nodes by a multi-dimensional scaling (MDS) approach, and the anchor nodes’ actual coordinates and target’s relative coordinates can be computed via coordinate transformation.
- Using online linear classifiers to filter spam emails28
The analysis and promising experimental results indicate that the Perceptron and Winnow are two very competitive classifiers for anti-spam filtering.
- 27
- 27
- 27
- An Extended Document Frequency Metric for Feature Selection in Text Categorization27
An extended DF method called TFDF is put forward which combines the Term Frequency (TF) factor and experimental results show that TFDF performs much better than the original DF method.
- CoBA: Collusive Backdoor Attacks With Optimized Trigger to Federated Learning25
It is demonstrated that colluding malicious participants can effectively execute backdoor attacks during the FL training process, exhibiting high sparsity and stealthiness, which means they can evade common defense methods with only a few attack iterations.
- A Practical Approach to Constructing a Geological Knowledge Graph: A Case Study of Mineral Exploration Data25
A novel framework that can extract the geological knowledge graph (GKG) from public reports relating to a modelling study is proposed and a geological knowledge representation model with three levels of “geological concepts- geological entities-geological relations” is explored to describe semantic units of geological knowledge and their logic relations.
- Learning Hierarchical Dynamics with Spatial Adjacency for Image Enhancement24
A DCP-guided hierarchical dynamic mechanism for image enhancement to adapt the model parameters and features from local to global as well as to keep spatial adjacency within the region to achieve state-of-the-art performance and generates visually pleasing images.
- Helicopter maritime search area planning based on a minimum bounding rectangle and K-means clustering24
Simulation experiment results show that the POS values obtained by the MBRK method are higher than those obtained by other methods, which proves that the MBK method can effectively support the planning of search areas and that K-means clustering improves the POS of search plans.
- A new hazard event classification model via deep learning and multifractal22
A novel deep learning model termed DLF through multifractal is presented to explore HaE classification where the motivation is that HaE can be naturally regarded as a kind of time series and results demonstrate that compared with other classifiers, DLF classifier performs better under metrics of precision, recall and F1-score.
- 22
- A Wolf in Sheep's Clothing: Practical Black-box Adversarial Attacks for Evading Learning-based Windows Malware Detection in the Wild21
MalGuise is a practical black-box adversarial attack framework that evaluates the security risks of existing learning-based Windows malware detection systems under the black-box setting and achieves a remarkably high attack success rate.
- 20
- Experiments in TREC 2004 Novelty Track at CAS-ICT.20
The main task in Novelty Track is to retrieve relevant sentences and remove duplicates from a document set given a TREC topic, which includes customized language parsing on original dataset, document retrieval, sentence relevance and novelty detection.
- 19
- 19
- Attentive Path Combination for Knowledge Graph Completion.19
This paper proposes a novel path-based relation inference model that learns entity pair representations with attentive path combination, and shows that the proposed model achieves better performance than state-of-the-art path- based relation inference methods.
- TREC-10 Experiments at CAS-ICT: Filtering, Web and QA.19
This paper tried to integrate different ranking methods to improve system performance and focused on question type identification, named entity tagging and answer matching for QA track of TREC-10.
- 18
- Outlier Detection over Sliding Windows for Probabilistic Data Streams18
The problem of detecting an outlier over a set of possible world instances is equivalent to the problem of finding the k-th element in its neighborhood and a dynamic programming algorithm is proposed to reduce the detection cost from O(2|R( e; d) to O(|k·R(e; d)|), where E is the d-neighborhood of e.
- Malware homology determination using visualized images and feature fusion17
This work sought to develop a homology determination method using the fusion of global structure features and local fine-grained features based on malware visualization, and proposes a dual-branch convolutional neural network, which features the opcode image and bytecode image as the final family classification basis.
- 17
- 17
- 16
- EFENet: Reference-Based Video Super-Resolution with Enhanced Flow Estimation16
The novel flow refinement module of EFENet refines the flow regarding the furthest frame using all the estimated flows, which leverages the global temporal information within the sequence and therefore effectively reduces the alignment errors.
- 16
- 16
- A Location Privacy Preserving Approach on Road Network16
A novel location privacy preserving approach, called cloaking cycle and forest (CCF for short), is proposed, which finds cycles and forest that satisfy certain conditions using breadth-first search in graph and can still protect the mobile user's location privacy when the road network contains one-way streets.
- 15
- Finding the PISTE: Towards Understanding Privacy Leaks in Vertical Federated Learning Systems15
This study exposes the fundamental susceptibility of VFL systems to privacy leaks, which arise from the collaboration between the server and clients during both training and testing, and proposes PISTE, a model-agnostic framework of privacy stealing attacks against VFL.
- Animating Images to Transfer CLIP for Video-Text Retrieval15
A novel image animation strategy to transfer the image-text CLIP model to video-text retrieval effectively by imitating the video shooting components and proposing a coarse to fine framework which consists of dual encoders for fast candidates searching and a cross-modality interaction module for fine-grained re-ranking.
- 15
- Minimizing Explanations for Missing Answers to Queries on Databases15
A novel explanation technique to avoid such unexpected explanations and a guarantee that the generalized explanation is minimum, and two algorithms by considering different kinds of modification object values for the attributes satisfying referential integrity constraint are proposed.
- 15
- 14
- KGBoost: A classification-based knowledge base completion method with negative sampling14
The new method, named KGBoost, adopts a modularized design and attempts to find hard negative samples so as to train a powerful classifier for missing link prediction, and outperforms state-of-the-art methods across most datasets.
- 14
- Entity-based query reformulation using wikipedia14
An automatic query reformulation approach based on entities detected in each query is presented to utilize semantics associated with entities for enhancing document retrieval and results show that significant improvement is possible by utilizing entity corresponding information.
- 12
- Binocular Video-Based Automatic Pixel-Level Crack Detection and Quantification Using Deep Convolutional Neural Networks for Concrete Structures12
A novel real-time crack detection and quantification method that leverages binocular vision and a lightweight deep learning model is proposed that can contribute to crack detection and quantification in practical use by being deployed on smart devices.
- 12
- 12
- Real-time and high-precision interrogation of a linearly chirped fiber Bragg grating sensor array based on dispersive time delay and optical pulse compression12
A novel technique to achieve real-time and high-precision interrogation of a linearly chirped fiber Bragg (LCFBG) grating sensor array based on dispersive time delay and optical pulse compression is proposed and experimentally demonstrated.
- A secure K‐automorphism privacy preserving approach with high data utility in social networks12
A secure and high utility privacy preserving model, called AK- secure, to prevent node identity attack, path length leakage, and edge leakage effectively is designed and a graph anonymous algorithm is proposed, which constructs an anonymous graph satisfying the AK-secure privacy preservingmodel, minimizing information loss and guaranteeing high data utility.
- A Comparative Study of Pseudo Relevance Feedback for Ad-hoc Retrieval12
Results show that a KL-divergence based DFR method, combined with the classical Rocchio's algorithm, has the best retrieval effectiveness out of the three methods studied in this paper.
- GKEEP: An Enhanced Graph‐Based Keyword Extractor With Error‐Feedback Propagation for Geoscience Reports11
An unsupervised algorithm, the graph‐based keyword extractor with error‐feedback propagation (GKEEP), that enhances graph-based keyword extraction approaches by using an error-feedback mechanism similar to the concept of backpropagation is proposed.
- Indexing Uncertain Data for Supporting Range Queries11
A novel structure called MRST is proposed to approximately capture the probability density function of uncertain object to provide uncertain object with strong pruning power and consume fewer space cost and an efficient algorithm is designed based on characters of MRST.
- CN-RMA: Combined Network with Ray Marching Aggregation for 3D Indoor Object Detection from Multi-View Images10
This paper introduces CN-RMA, a novel approach for 3D indoor object detection from multi-view images that leverages the synergy of 3D reconstruction networks and 3D object detection networks, where the reconstruction network provides a rough Truncated Signed Distance Function (TSDF) and guides image features to vote to 3D space correctly in an end-to-end manner.
- The discrimination of tectonic settings using trace elements in magmatic zircons: A machine learning approach10
EXtreme Gradient Boosting could provide a high-dimensional discriminative approach to distinguish tectonic settings and its results for accuracy, precision, recall, and F1 score are impressive.
- A Deep Convolutional Neural Network Based Risk Identification Method for E-Commerce Supply Chain Finance10
A multigroup experimental study shows that the accuracy of supply chain finance risk identification using deep convolutional neural network models can reach 95.36%, which demonstrates the effectiveness of the proposed method by providing better performance compared to traditional BP and SVM networks.
- CCUBI: A cross‐chain based premium competition scheme with privacy preservation for usage‐based insurance10
This study proposes CCUBI, a cross‐chain‐based premium competition scheme with privacy preservation for intelligent UBI systems, and introduces tamper‐resistant blockchain and smart contracts to construct credible insurance mechanisms.
- 10
- H-MRST: A Novel Framework For Supporting Probability Degree Range Query Using Extreme Learning Machine10
This paper proposes a new query called PDR (short for probabilistic degree range) query to substitute the traditional prob-range query, and proposes an ELM-based “adapter” to construct the lightweight structure for uncertain data in a more general manner, and designs the GO-ELM algorithm for answering PDR query.
- 10
- 10
- 10
- Query Performance Prediction for Information Retrieval Based on Covering Topic Score10
Experimental results show that CTS consistently performs better than previous state-of-the-art methods for query performance prediction including clarity score and robustness score, and is also shown to have very low computational complexity, meaning that it can be practical for real applications.
- Information retrieval oriented word segmentation based on character associative strength ranking10
This paper presents a novel, ranking-style word segmentation approach, called RSVM-Seg, which is well tailored to Chinese information retrieval (CIR), and is able to eliminate overlapping ambiguity much more effectively, compared to the currentword segmentation methods.
- Robust Backdoor Attacks on Object Detection in Real World9
A variable-size backdoor trigger to adapt to the different sizes of attacked objects, overcoming the disturbance caused by the distance between the viewing point and attacked object and a malicious adversarial training, enabling the backdoor object detector to learn the feature of the trigger with physical noise.
- A high-effective multitask surface defect detection method based on CBAM and atrous convolution9
Experiments show the superior accuracy and real-time performance of the proposed model when compared with current mainstream surface defect detection methods and indicate its wide applicability in the detection of surface defects in industrial products.
- Exploring Security Vulnerabilities of Deep Learning Models by Adversarial Attacks9
A visual similar word replacement (VSWR) algorithm is proposed to generate adversarial examples against textual analysis models to explore the security vulnerabilities of deep learning models designed for textual analysis.
- Resolving Entity Morphs based on Character-Word Embedding9
This paper focuses on morph resolution: given a morph, figure out the entity that it really refers to after analyse the common characteristic of morphs and target entities from cross-source corpora, and proposes a framework based on character-word embeddings and radical-character-wordembeddings to rank target candidates.
- A retrospective study of female reproductive parameters in the Kunming dog9
Bitch parity influenced litter size, and the gestation length and birth weight of the puppies were negatively affected by litter size), which confirms that Kunming dogs are non-seasonal breeders.
- Computing and applying topic-level user interactions in microblog recommendation9
This work uses topic model to analyze users' interactive behaviors and measure the topic-specific relationship strength, then incorporates the relationship factor into the matrix factorization framework, which can achieve better performance on the MAP and NDCG evaluation measures, and have better interpretability for the recommended results.
- 9
- 9
- 8
- 8
- 8
- 8
- 8
- Robust Network Architecture Search via Feature Distortion Restraining8
This work defines a network vulnerability metric based on feature distortion as a constraint in the search process to obtain a robust network against adversarial attacks and proposes Robust Network Architecture Search (RNAS), a novel algorithm to solve this optimization.
- Improving Web Document Clustering through Employing User-Related Tag Expansion Techniques8
A user-related tag expansion method, which incorporates additional useful tags into the original tag document by utilizing user tagging data as background knowledge, and can alleviate topic drift in expansion, which indicates that tags could be a better resource for the clustering task.
- ICTIR Subtopic Mining System at NTCIR-9 INTENT Task.8
This paper describes the approaches and results of the Chinese subtopic mining system for the NTCIR-9 INTENT task, which first finds out the related queries from query logs, then group them into different clusters using a frequent term-set based clustering algorithm.
- 7
- A Blockchain-Based IoT Cross-Domain Delegation Access Control Method7
The proposed blockchain-based IoT cross-domain delegation access control method (CDDAC) has faster delegation verification speed and higher decision-making efficiency than existing work, demonstrating the lightweight and scalability of the method.
- A Time-Aware Language Model for Microblog Retrieval.7
A ranking algorithm with temporal information based query that suggested that HTLM models are more effective for Microblog search and mix model further improved compared with the single model.
- User-aware caching and prefetching query results in web search engines7
Experimental results show that the proposed User-Aware Cache approach can achieve hit ratios better than state-of-the-art methods.
- User-Related Tag Expansion for Web Document Clustering7
A user-related tag expansion method, which incorporates additional useful tags into the original tag document by utilizing user tagging as background knowledge, and can alleviate the topic drift in expansion, especially for those topic-specific documents.
- Using Global Statistics to Rank Retrieval Systems without Relevance Judgments7
A novel method using global statistics to rank retrieval systems without relevance judgments, which outperforms all early attempts and is adjustable for different effectiveness measurements, e.g. MAP, precision at n, and so forth.
- From SRME or wave-equation extrapolation to SRME and wave-equation extrapolation7
Among the three techniques, hybrid adaptive subtraction takes full advantage of the two multiple prediction techniques, and emerges as champion.
- 7
- When Internet companies morph7
The study provides academic researchers and executives with an overview of industry strategic mutation patterns and provides executives with a process analysis for identifying and evaluating their own strategies in a way that is essential for success in the highly volatile Internet economy.
- Beyond Single Images: Retrieval Self-Augmented Unsupervised Camouflaged Object Detection6
A Clustering-thenRetrieval (CR) strategy, where coarse masks are first generated through clustering, facilitating subsequent histogrambased image filtering and cross-category retrieval to produce high-confidence prototypes, is introduced.
- Public concerns analysis and early warning of Mpox based on network data platforms—taking Baidu and WeChat as example6
The integrated use of big data offers a platform for rapid understanding public concerns and early warning signs of emerging infectious diseases such as Mpox as well as its potential early warning role.
- Neighbor-Enhanced Representation Learning for Link Prediction in Dynamic Heterogeneous Attributed Networks6
This work designs a neighbor-enhanced dynamic heterogeneous attributed network embedding method (NeiDyHNE) for link prediction that encodes the attributes and neighborhood structure of nodes as well as the evolutionary features of the dynamic network.
- PnA: Robust Aggregation Against Poisoning Attacks to Federated Learning for Edge Intelligence6
This work proposes a robust aggregation mechanism, called Parts and All (PnA), to protect the global model of FL by filtering out malicious local model updates throughout the detection of poisoning attacks at layers of local model updates.
- 6
- 6
- A Study on Particle Swarm Algorithm Based on Restart Strategy and Adaptive Dynamic Mechanism6
The RDS-PSO algorithm proposed in this paper solves the problem of falling into local extremums and precocious puberty, significantly improves the optimization, speed, and effectiveness of the path, and the simulation experiments also show that the algorithm has good robustness and generalization.
- SenCS6
A context-based indoor proximity verification scheme, called SenCS, to enable real-time autonomous access for mobile devices, utilizing the available heterogeneous sensors at the user side and at the room side, and exploiting the time intervals between successively human actions to create heterogeneous contextual fingerprints (HCF) at a millisecond level.
- 6
- Polyphone Disambiguation in Mandarin Chinese with Semi-Supervised Learning6
This paper proposes a novel semi-supervised learning (SSL) framework for Mandarin Chinese polyphone disambiguation that can potentially leverage unlimited unlabeled text data and explores the effect of various proxy labeling strategies including entropy-thresholding and lexicon-based labeling.
- Incorporating dynamic factors for improving a GIS‐based solar radiation model6
This research presents a probabilistic procedure for estimating the intensity and direction of emplacing carbon dioxide in the atmosphere over a large area using a simple, scalable, and scalable method.
- 6
- Birth of puppies of predetermined sex after artificial insemination with a low number of sex‐sorted, frozen–thawed spermatozoa in field conditions6
Preliminary information indicated that normal puppies of predicted sex can be produced with low numbers of sorted cryopreserved dog spermatozoa at a farm level, making sperm-sexing technology potentially applicable for elite breeding units.
- A Wordnet-Based Approach to Feature Selection in Text Categorization6
In this method, word tendency, which takes related words into consideration, is used to select best terms and the best performance is usually achieved in fewer features.
- Design and Analysis of a Channel Access Protocol for WDM-Based Single-Hop Optical Networks*6
Through numerical results and simulation studies, it is shown that the proposed protocol achieves better channel utilization and incurs lower packet delays.
- 5
- 5
- Adversarial perturbation denoising utilizing common characteristics in deep feature space5
The defense method can eliminate multiple adversarial perturbations, significantly enhancing adversarial robustness compared to previous state-of-the-art methods and can be plug-and-play for various classification models, which demonstrates the generalizability of the defense method.
- 5
- 5
- Hessian-Free Second-Order Adversarial Examples for Adversarial Learning5
This work crafts second-order adversarial examples and uses them to train DNNs and proposes an approximation method through transforming the problem into an optimization in the Krylov subspace, which remarkably reduce the computational complexity to speed up the training procedure.
- Visually imperceptible adversarial patch attacks5
The main idea is to create a contributing feature region (CFR) of an image by simulating the human attention mechanism and then add perturbations to CFR, and a soft mask matrix is designed on the basis of activation map to represent the contributions of each pixel in CFR.
- Nebula-I: A General Framework for Collaboratively Training Deep Learning Models on Low-Bandwidth Cloud Clusters5
A general framework for collaboratively training deep learning models over remote heterogeneous clusters, the connections between which are low-bandwidth wide area networks (WANs), and experiments demonstrate that the proposed framework could substantially maximize the training efficiency while preserving satisfactory NLP performance.
- An end-to-end dehazing network with transitional convolution layer5
A simple but effective dehazing network by learning the residual image between the hazy image and haze-free image is proposed, and a novel transitional convolution structure is constructed in this network, which contributes to utilizing shallow structure information to enhance the final residual image.
- 5
- Spatiotemporal Breast Mass Detection Network (MD-Net) in 4D DCE-MRI Images5
This work aims to leverage recent deep learning techniques for breast lesion detection and proposes the Spatiotemporal Breast Mass Detection Networks (MD-Nets) to detect the masses in the 4D DCE-MRI images automatically.
- Crowdsourced Indoor Localization for Diverse Devices with RSSI Sequences5
This work presents an efficient localization method with crowdsourced users’ RSSI sequences, and proposes a bit encoder method to prune the RSSIs that cannot be the result when localizing an query user.
- 5
- Stochastic Fractal Search Algorithm for 3D Protein Structure Prediction5
Experiments show that the proposed method outperforms other algorithms on the accuracy of calculating the protein sequence energy value, which is turned out to be an effective way to analyze protein structure.
- 5
- 5
- 5
- A Category-integrated Language Model for Question Retrieval in Community Question Answering5
A novel Category-integrated Language Model (CLM) is proposed which views category-specific term saliency as the Dirichlet hyper-parameter that weights the parameters of LM and can significantly outperform the previous work that incorporates the category information into LM externally at the word level or at the document level.
- Author Model and Negative Feedback Methods on TREC 2011 Microblog Track.5
Experimental results show that the ICTIR methods could improve the retrieval performance greatly and the basic query likelihood model with smoothing is the fundamental method.
- 5
- 5
- A General Framework of Feature Selection for Text Categorization5
A novel feature selection framework called Distribution-Based Feature Selection (DBFS) based on distribution difference of features is proposed, which generalizes most of the state-of-the-art feature selection methods including OCFS, MI, ECE, IG, CHI and OR.
- GJM-2: A Special Case of General Jelinek-Mercer Smoothing Method for Language Modeling Approach to Ad Hoc IR5
Experimental results show that using GJM-2 for the language modeling approach can achieve better retrieval performances than the existing three popular methods both on short and long queries.
- 4
- 4
- 4
- Towards the Desirable Decision Boundary by Moderate-Margin Adversarial Training4
A novel adversarial training scheme that aims to learn a moderate-inclusive decision boundary, which means that the margins of natural examples under the decision boundary are moderate, which achieves high natural accuracy and robustness under both black-box and white-box attacks.
- Rethinking Reinforcement Learning based Logic Synthesis4
A new RL-based method that can automatically recognize critical operators and generate common operator sequences generalizable to unseen circuits is developed that achieves a good balance among delay, area and runtime, and is practical for industrial usage.
- Practical approximate indoor nearest neighbour locating with crowdsourced RSSIs4
A graph with interval weights, called I-graph, is proposed, which can connect the RSSIs and represent the topology of indoor space and construct a search tree index D-tree, which could search the nearest neighbour objects efficiently, and a novel distance metric for RSSI is proposed.
- 4
- A system-aware optimized data organization for efficient scientific analytics4
This study proposes a system-aware optimized data organization strategy that can organize data blocks of multidimensional scientific data efficiently based on simulation output and the underlying storage systems, thereby enabling efficient scientific analytics.
- Exploring categorization property of social annotations for information retrieval4
A novel TR-LDA model is presented to take annotations as an additional source for generating document contents apart from the document itself, and strategies for representing and weighting the categorization property are provided.
- 4
- Enhancing Keyword Search in Relational Databases Using Nearly Duplicate Records.4
The motivation is discussed and a system, RSEARCH, is presented, to show challenges in supporting keyword search using nearly duplicate records and key techniques including identifying nearly duplication records and generating results efficiently.
- An Axiomatic Approach to Exploit Term Dependencies in Language Model4
This paper derives an axiomatic dependence model by defining several basic desirable constraints that a retrieval model should meet and shows that this model significantly and robustly improves retrieval accuracy over the baseline (unigram model) in verbose queries and achieves better performance than some state-of-art dependence models.
- A Survey of Web Search Query Intention Classification4
The background of query classification, its key techniques, the classification algorithms and the evaluation methods are presented, and the problems and challenges in query intention classification are outlined.
- An Analysis of Question Processing of English and Chinese for the NTCIR 5 Cross-Language Question Answering Task4
Using the NTCIR 5 Cross-Language Question Answering (CLQA) question test set, it is demonstrated that the accuracy of deep question analysis is dependent on the quantity and suitability of the available linguistic resources.
- Example-based phrase translation in Chinese-English CLIR4
An example-based phrase translation method in a Chinese to English cross-language information retrieval (CLIR) system that can generate much more accurate query translations than dictionary-based and common MT-based methods is proposed.
- Anisotropic Gauss Reconstruction and Global Orientation with Octree‐based Acceleration3
This work considers the anisotropic Laplace equation to derive the anisotropic Gauss formula and applies it to surface reconstruction, called “anisotropic Gauss reconstruction”, and proposes an octree‐based acceleration algorithm to achieve a space complexity of O(N) and a time complexity of O(NlogN).
- 3
- 3
- 3
- Order-aware Interactive Segmentation3
This work proposes OIS: order-aware interactive segmentation, where the relative depth between objects is explicitly encoded into order maps, and introduces a novel order-aware attention, where the order maps seamlessly guide the user interactions to attend to the image features.
- Defending Privacy Inference Attacks to Federated Learning for Intelligent IoT with Parameter Compression3
Inspired by the idea of gradient compression, a defense method called Federated Learning Parameter Compression (FLPC) is proposed which can reduce the sharing of information for privacy protection and prevents attackers from recovering the private information of victims while maintaining the accuracy of the global model.
- Research on intelligent customs declaration generation in Guangdong-Hong Kong cross-border road cargo clearance3
This research applied to the China International Trade Single Window (Shenzhen) and significantly improved the efficiency and accuracy of customs declaration in GD and HK.
- 3
- AsyncET: Asynchronous Learning for Knowledge Graph Entity Typing with Auxiliary Relations3
A method adopting an Asynchronous learning scheme for Entity Typing, named AsyncET, which updates the entity and type embeddings alternatively to keep the learned entity embedding up-to-date and informative for entity type prediction is proposed.
- 3
- 3
- 3
- Subtle Hand Action Recognition in Factory Based on Inertial Sensors3
An operation action recognition method based on two wearable devices containing inertial sensors is proposed, using a new idea that using the acceleration and gyroscope data, while segment the action in the same time to improve the precision for the segmentation technology.
- Towards Understanding URL Resources in Recent Sina Weibo3
A corpus which contains the dissemination and classification information about URLs in Sina Weibo is constructed and it is found that verified users tend to publish about twice the amount of URLs as non-verfied users; Video URLs are more easily to disseminate in Sina weibo.
- 3
- Query Reformulation Based on User Habits for Query-by-Humming Systems3
A novel way of query reformulation for QBH search engine by modeling the user’s humming habits and proposed a user-humming-model based on Hidden Markov model, which improves the quality of the user humming and improves the effect of information retrieval.
- 3
- 3
- Automatic Refinement of Keyword Annotations for Web Image Search3
This paper proposes a framework to utilize the visual content, the textual context, and the semantic relations between keywords to refine the image annotation and demonstrates the effectiveness of the proposed method.
- Tree-like database design model for expansible application3
The new model is proved to be more suitable for the complicated and changeable database systems whose applications are possible to expand, by giving an example of a company's product information system model and comparing it with the traditional design model in ERP software.
- CAS-ICT at TREC 2005 SPAM Track: Using Non-Textual Information to Improve Spam Filtering Performance3
This paper introduces the work in the TREC2005 SPAM track, and results show that Winnow outperforms Naive Bayes and the multi-feature model outperforms structure based model.
- 2
- 2
- Deepfake Video Traceability and Authentication via Source Attribution2
A novel training‐free deepfake detection framework based on the interpretable inherent source attribution that provides a robust and efficient solution for deepfake video authentication and source attribution, thus, making it highly adaptable to real‐world scenarios.
- 2
- 2
- 2
- 2
- 2
- A Novel Privacy‐Preserving Mobile‐Coverage Scheme Based on Trustworthiness in HWSNs2
A novel privacy-preserving mobile coverage scheme based on trustworthiness that can efficiently mitigate some malicious attacks such as eavesdropping and pollution and optimize the coverage of hybrid wireless sensor networks (HWSNs) at the same time is proposed.
- Processing Continuous k Nearest Neighbor Queries in Obstructed Space with Voronoi Diagrams2
With the emergence and growing popularity of and location-based service (LBS) technologies, the continuous k nearest neighbor query in obstructed space is becoming a very important service.
- Towards Imperceptible Adversarial Image Patches Based on Network Explanations.2
A novel method to generate local region perturbations to find the contributing feature regions (CFRs) of images based on network explanations for perturbation, using a soft mask matrix to represent the CFRs for finely characterizing the contributions of each pixel.
- Temporal Smoothing: Discriminatively Incorporating Various Temporal Profiles of Queries2
By using the overall temporal distribution of documents to smooth the temporal profile of a given query, the estimated temporal query model can be used in query likelihood retrieval model as an unified solution.
- 2
- Research and Application of Anomaly Detection of Industrial Control System Based on Improved Zoe Algorithm2
An improved Zoe algorithm is proposed that can achieve higher detection rate and lower false positive rate of anomaly detection in industrial control systems.
- Event Related Document Retrieval Based on Bipartite Graph2
A novel bipartite graph model is proposed to describe an event, where one bipartition represents event type and the other represents the event specific information, and this model is integrated into the query model for more accurate query representation.
- 2
- 2
- 2
- 2
- 2
- 2
- Prediction of rotary machinery degradation status based on vibration data using back propagation (BP) neural network2
The results showed that the Improved Variable Learning Rate (IVLR) model only required eight min, which is half of the Traditional Variable Learning rate (TVLR) processing time to converge to the global minimum, and showed better nonlinear approximation ability.
- Improving Text Search on Hybrid Data2
This paper describes a new method about improving text search using structured data and builds the uniform problem model, which is the first approach adopting the mutual information of feature words to qualify the relevance (similarity) between two texts.
- 2
- 2
- Semi-supervised learning and opinion-oriented information extraction2
Three types of semi-supervised learning algorithms are focused on, namely, self-training, co- training, and graph-based methods, which are used to solve sentence subjectivity classification, contextual polarity recognition, and opinion entity identification.
- 2
- 1
- Towards Ship License Plate Recognition in the Wild: A Large Benchmark and Strong Baseline1
A large dataset, called SLP34K, which consists of 34,385 images collected by an intelligent traffic surveillance system, is built, which is carefully manually annotated with text labels and attributes, and presents high data diversity by multiple installation locations and long capturing period of the cameras.
- TaP: A Taxonomy-Guided Framework for Automated and Scalable Preference Data Generation1
This work uses TaP-generated datasets to perform supervised and preference fine-tuning on multiple LLMs, and demonstrates that LLMs trained on TaP-generated datasets outperform those trained on existing open-source datasets.
- 1
- Learn A Flexible Exploration Model for Parameterized Action Markov Decision Processes1
This work proposes a model-based RL (MBRL) algorithm, FLEXplore, which learns a parameterized-action-conditioned dynamics model and employs a modified Model Predictive Path Integral control, and carefully design the dynamics loss function and reward smoothing process to learn a loose yet flexible model.
- 1
- Adversarial training in logit space against tiny perturbations1
The results show that the proposed novel EAE adversarial training not only shortens the training time, but also enhances the robustness of model and has less impact on the accuracy of clean examples than the existing state-of-the-art methods.
- 1
- 1
- BIT-Xiaomi’s System for AutoSimTrans 20221
The BIT-Xiaomi simultaneous translation system for Autosimtrans 2022 simultaneous translation challenge is described, and wait-k is employed to train prefix-to-prefix translation models and streaming chunking is integrated to detect boundaries as the source streaming read in.
- 1
- 1
- 1
- Construction of Quality Monitoring System for Online Courses of Navigation Majors1
A "four-in-one" evaluation system model of online course quality is constructed, and a four-level linkage monitoring system is established to build a course quality spiral improvement model and ensure the substantial equivalence of online teaching and offline teaching.
- 1
- 1
- Improving Retrieval Quality Using PRF Mechanism from Event Perspective1
Empirical evaluations based on TREC-TS 2015 dataset and standard benchmarks, and state-of-the-art pseudo feedback methods, demonstrate the effectiveness of the proposed EO-PRF mechanism in event-oriented retrieval.
- 1
- 1
- 1
- Evaluation of the Application Benefit of Meteorological High Performance Computing Resources1
The concept of the utility value B and index compliance rate E of the meteorological high performance Computing system are presented and the evaluation process, evaluation index and calculation method of the high performance computing resource application benefits are introduced.
- 1
- A Set-Based Training Query Classification Approach for Twitter Search1
A set-based training query classification approach that estimates a training query’s quality by taking its usefulness in combination with other training queries into consideration is proposed.
- Prevalence of Anomalies in Real World Sensor Network Deployments1
This chapter discusses a few sensor network deployments to emphasize the need for an anomaly detection mechanism in the presence of failing nodes, frequent network failures, and/or potentially malicious network behaviors.
- Multi-constrained Optimal Path Search Algorithms1
This paper proposes an exact algorithm to find the optimal path in relatively small data sets, and presents a heuristic algorithm which is efficient and scalable to large data sets.
- Research on GPS Signal Capturing1
The compatible development trend with the coexistence of multiple systems requires additional performances of the receiver and the Double Dwell Sliding Correlation Method is used to capture the satellite data in the MATLAB.
- 1
- Switched optimization algorithm of piecewise contractive path1
The paper analyzed the rules of switched control on the contractive path and found the least length and weighted average contractive paths and peak value control algorithms that showed the stability and convergence under different optimal algorithms.
- Viewpoint Selection Using Hybrid Simplex Search and Particle Swarm Optimization for Volume Rendering1
A novel method of viewpoint selection using the hybrid Nelder-Mead (NM) simplex search and particle swarm optimization (PSO) to improve the efficiency and the intelligent level of volume rendering is proposed.
- Study of BGP secure scheme based on divide and conquer strategy1
Performance evaluation results show that SA-BGP can be implemented efficiently and the incurred overhead, in terms of time and space, is acceptable in practice.
- A Query Result Caching and Prefetching Approach Based on Query Characteristics1
A query results caching and prefetching approach based on query characteristics is proposed and can achieve 3.5% to 8.45% increase for all requests as compared with state-of-the-art methods.
- Study on Dynamic Balance in Main Driving Mechanism of Cold Rolling Mill Based on the Particle Swarm Optimization Algorithm1
The simulation results indicate that the proposed method could eliminate the reluctant evaluations and interactions remarkably, thus improves the application's performance.
- 1
- Application of Preliminary Hazard Analysis in Rapid Transit Systems1
Insight is given into safety of the rapid transit system by means of improved preliminary hazard analysis (PHA) and results indicated that some sections of the Rapid Transit System Phase II failed to meet the safety requirements and are exposed to hazards.
- 1
- The EQ3 format of seismic catalogue and its implementation method in SQL Server Database1
The definition of the EQ3 format in C program language is introduced, and the comparison of EQ3 and other used formats such as EQT,EQ2 WKF and SHK is made in this paper.
- 1
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- Design of welding machine for automobile parts–
The hardware part mainly includes the STC12C5A32AD single chip microcomputer as the control core, which uses the key input and the sensor to complete the system information collection, MCU control relay, solenoid valve on and off to achieve welding power, pneumatic components and other executive control.
- An Efficient Evolutionary Deep Learning Framework Based on Multi-source Transfer Learning to Evolve Deep Convolutional Neural Networks–
A new evolutionary computation based framework is proposed to efficiently evolve CNNs without compromising the classification accuracy, which shows the experimental results show the proposed method acquires good CNNs faster than 15 peer competitors within less than 40 GPU-hours.
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- Development of Operational Technology for Meteorological High Performance Computing–
The performance calculation business of the meteorological department will usher in a better tomorrow, and high-performance computing technology and meteorological numerical forecasting applications are increasingly integrated, and continue to innovate and develop.
- –
- –
- –
- –
- –
- –
- A self-adaptive anti-vibration pipeline-filtering algorithm–
Experimental results show that the algorithm can overcome the problem of interframe vibration of sequence images, thus realizing accurate detection of small dim maritime targets.
- Obtaining K‐obfuscation for profile privacy in social networks–
The definition of profile privacy leakage is formally presented, which is a newly identified privacy leakage in social networks, and a novel safe vertex-profile mapping mechanism, named as k-mapping is proposed.
- –
- –
- APRED:RED Improved Algorithm Based on Avg and P–
Simulation results show that the improved APRED algorithm, which dynamically adjusts the avg and non-linear modifies P, can get a better average throughput in the bottleneck link and has lower packet loss rate in router.
- –
- –
- –
- –
- –
- VC Based Design and Implementation of Database Programming–
The characteristics of VC database programming are analyzed, the VC environment of the two main database development model is compared, and the details of ADO programming model example are given.
- Study on New Type Agriculture Information Service Mode–
"148" service mode was put forward, which provides a new service mode for local modern agriculture upgrading and is a leading technology in the province and is worth for farmers' application.
- On Electromagnetic Trace in Perspective of Criminal Investigation–
As a kind of important investigation method, the tracing of electromagnetic information has been widely used in the criminal investigation, and has been playing a more and more important role.
- –
- Anomalous traffic classification based on maximum entropy model–
The results of experiment show that the maximum entropy model has better classification efficiency, especially under small size of training data set.
- –
- –
- Nautical nuclear chemical and biological security threat and collective protection of surface ship–
By comparing the collective protection ability of foreign naval main active surface ships, the technology essentials and the ability necessary of surface ship NCB collective protection are analyzed and shows clearly the further development direction of collective protection designment.
- Design of Meteorological Computing Resources Management System–
A management system of meteorological computing resources is proposed that implements the functions for resources charging, allocation and statistics functions by database technology.
- Collaborative Platform for the Preparation of Dispatch Daily and Shift Plans Based on Blackboard Model–
It has been proven that the platform for the preparation of dispatch daily and shift plans is an effective way to solve the problem of cooperative work between multi-type dispatchers.
- –
- Research on extension of information services under scientific computing grid environment–
Experimental results show the feasibility of the design scheme, and the management cost is within reasonable limits, which meets well the demands for building scientific-computing-oriented grid environment.
- Design and implementation of security monitor system for USB storage device–
This paper introduced the existing USB storage device monitoring technology, and analyzed the main problem, and presented a design scheme of security monitoring by means of the disk lower filter based on the architecture of Win32 Driver Model(WDM).
- –
- Implementation of Buffer Manager in Ubase Database System–
Love/Hate algorithm is finally chosen instead of LRU algorithm to implement the buffer management of Ubase database system, which reduces the cost of system and improves the performance of U Base database system.
- Construction emphases of the basic geographical information platform on a province area–
This paper discussed the key technology and system building thoughts in the Chongqing Geomatics Center building of basic geographical information platform that included data standard, data check, database, management and Web Service.
- Research and Application on MVC Framework in ASP.NET–
This paper completes the applied analysis of a basic project of ASP.NET MVC framework and provides a new Model-View-Controller framework on top of the existing ASP.
- –
- –
- Analysis on SQL Server Database Deployment Way–
This paper analyzes each kind of SQL Server database deployment way, complete the deployment of the database by SQL server database enterprise tool or "Data Publishing Wizard" in Visual Studio.
- –
- Utilizing phrase based semantic information for term dependency–
This work uses Wikipedia to separate important and less important term dependencies, and treats them accordingly as features in a linear feature-based retrieval model, and compares the method with a Markov Random Field model on four TREC document collections.
- Application of the Wireless Transmission Technology in the Train Safety Monitoring System–
A wireless data-transmission scheme that adopts GPRS network is presented and the result shows that the wireless transmission technique applied in the train safety monitoring system is reliable and practicable.
- Distributed WEB Information Retrieval Based on Link Partition–
The efficiency experimental results indicate that the LIBCA plus BloomFiltern achieves a high system performance and it can deal with large data-set and it is shown that at P@10 the results of link partition based distributed WEB information retrieval are equal or better than that of centralized retrieval.
- An investigation of question translation for English-Chinese cross-language question answering–
An evaluation of English–Chinese CLQA comparing question translation using standard machine translation systems and extended translation incorporating web mining to enhance the translation dictionary against a baseline of monolingual Chinese QA shows that the noun phrase recognition and translation techniques lead to a significant improvement inCLQA effectiveness.
- –
- –
- –
- –
- –
- A new method of accelerating the speed and the data access–
This article gives a detatiled introduction about how to deal with such kind of questions with notebook by using data interface constructed with C++, which improves accessing database.
- A Brief Discussion on the Internet Safty–
The safty and the proctection of computer using in different aspects are discussed and the importance of the internet safty is get to know and then take some protective measures according to the specific problems.
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.