Is this you? Claim this profile to correct it, add a bio and choose the work people see first.
Claim this profileWorks402 from public data
- A composite kernel to extract relations between entities with both flat and structured features286
This study illustrates that the composite kernel can effectively capture both flat and structured features without the need for extensive feature engineering, and can also easily scale to include more features.
- Code Bad Smells: a review of current knowledge213
A systematic literature review of 319 papers published since Fowler et al. identified Code Bad Smells suggests that there is little evidence currently available to justify using Code Good Smells.
- Large Language Model Cascades with Mixture of Thoughts Representations for Cost-efficient Reasoning195
This paper proposes a cascade pipeline that follows the intuition that simpler questions can be addressed by a weaker but more affordable LLM, whereas only the challenging questions necessitate the stronger and more expensive LLM.
- Joint Multi-modal Aspect-Sentiment Analysis with Auxiliary Cross-modal Relation Detection168
This paper is the first to jointly perform multi-modal ATE (MATE) and multi- modal ASC (MASC), and it proposes a multi-Modal joint learning approach with auxiliary cross-modAL relation detection for multi- Modal aspect-level sentiment analysis (MALSA).
- RaSa: Relation and Sensitivity Aware Representation Learning for Text-based Person Search150
A Relation and Sensitivity aware representation learning method that outperforms existing state-of-the-art methods by 6.94%, 4.45% and 15.35% in terms of Rank@1 on CUHK-P EDES, ICFG-PEDES and RSTPReid datasets, respectively.
- A generation model to unify topic relevance and lexicon-based sentiment for opinion retrieval132
A novel generation model that unifies topic-relevance and opinion generation by a quadratic combination is proposed and demonstrates that in the opinion retrieval task, a Bayesian approach to combining multiple ranking functions is superior to using a linear combination.
- Do users rate or review?: boost phrase-level sentiment labeling with review-level sentiment classification125
This paper attempts to bridge the gap between phrase-level and review/document- level sentiment analysis by leveraging the results given by review-level sentiment classification to boost phrase- Level sentiment polarity labeling in contextual sentiment lexicon construction tasks, using a novel constrained convex optimization framework.
- Aspect Sentiment Classification with both Word-level and Clause-level Attention Networks105
The need for incorporating the importance degrees of both words and clauses inside a sentence is highlighted and a hierarchical network with both word-level and clause-level attentions to aspect sentiment classification is proposed.
- Code Summarization with Structure-induced Transformer102
ructure-induced Transformer is proposed, which encodes sequential code inputs with multi-view structural clues in terms of a newly-proposed structure-induced self-attention mechanism, and helps achieve new state-of-the-art results on benchmarks.
- Understanding User Experience in Large Language Model Interactions97
This study develops a taxonomy of 7 user intents in LLM interactions, grounded in analysis of real-world user interaction logs and human verification, and conducts a user survey to gauge their satisfaction with LLM services, encompassing usage frequency, experiences across intents, and predominant concerns.
- Emotion Tokens: Bridging the Gap among Multilingual Twitter Sentiment Analysis97
This paper focuses on sentiment analysis for Twitter messages (tweets) by the analysis of emotion tokens, including emotion symbols, irregular forms of words and combined punctuations, which become a useful signal for sentiment analysis on multilingual tweets.
- Overview of the NTCIR-9 INTENT Task90
This is an overview of the NTCIR-9 INTENT task, which comprises the Subtopic Mining and the Document Ranking subtask, and describes the subtasks, data and evaluation methods, and reports on the results.
- VideoVista: A Versatile Benchmark for Video Understanding and Reasoning80
This work presents VideoVista, a video QA benchmark that integrates challenges across diverse content categories, durations, and abilities and presents an automatic data construction framework, leveraging powerful GPT-4o alongside advanced analysis tools to enhance the capabilities of video-related LMMs (Video-LMMs).
- Microblog Sentiment Analysis with Emoticon Space Model78
The emoticon space model (ESM) is proposed that leverages more emoticons to construct word representations from a massive amount of unlabeled data and helps identify subjectivity, polarity, and emotion in microblog environments.
- White Blood Cells Classification with Deep Convolutional Neural Networks72
The necessary step in the diagnosis of leukemia by the attending physician is to classify the white blood cells in the bone marrow, which requires the attending doctor to have a wealth of clinic experience.
- Overview of the NTCIR-11 IMine Task72
An overview of the NTCIR IMine task, which is a core task of NTCir-11 and also a succeeding work of INTENT@NTCIR-9 and INTENT2@N TCIR-10 tasks, and describes the task details, annotation of results, evaluation strategies and then the official evaluation results.
- 70
- From Skimming to Reading: A Two-stage Examination Model for Web Search70
This study designs an experimental search engine to collect both the user's feedback on their examinations and the eye-tracking/click-through data and proposes a two-stage examination model, which suggests that users make judgments according to different signals in different stages.
- Aspect Sentiment Classification with Document-level Sentiment Preference Modeling69
A Cooperative Graph Attention Networks (CoGAN) approach for cooperatively learning the aspect-related sentence representation and the importance of the document-level sentiment preference information to ASC is justified and the effectiveness of the approach capturing such information is demonstrated.
- Overview of the NTCIR-10 INTENT-2 Task.66
This paper provides an overview of the NTCIR-10 INTENT-2 task (the second INTENT task), which comprises the Subtopic Mining and the Document Ranking subtask, and describes the subtasks, data and evaluation methods.
- Expressive and Secure Searchable Encryption in the Public Key Setting (Full Version).65
This paper presents an expressive and secure asymmetric searchable encryption (ESASE) scheme, which is the first that simultaneously supports conjunctive, disjunctive and negation search operations, and analyzes the efficiency and proves it is secure in the standard model.
- Towards Reasoning in Large Language Models via Multi-Agent Peer Review Collaboration64
This work introduces a multi-agent collaboration strategy that emulates the academic peer review process and shows that the collaboration approach delivers superior accuracy across all ten datasets compared to existing methods.
- 64
- A multicenter clinical AI system study for detection and diagnosis of focal liver lesions62
The evidence suggests that LiAIDS can serve as a routine diagnostic tool and enhance the diagnostic capabilities of radiologists for liver lesions, and with the assistance of LiAIDS, the diagnostic accuracy of all radiologists improved.
- Grey Wolf Optimization algorithm based on Cauchy-Gaussian mutation and improved search strategy61
The paper proposed a Grey Wolf Optimization algorithm based on Cauchy-Gaussian mutation and improved search strategy (CG-GWO), introduced to increase the population diversity of the leader wolves and improve the global search ability of the algorithm.
- An Analysis of BERT in Document Ranking55
A state-of-the-art BERT-based ranking model with focus on its attention mechanism and interaction behavior is investigated and it is indicated that it is possible to transform BERT into a more efficient representation-focused model.
- Chaotic arithmetic optimization algorithm54
The experimental results reveal that CAOA can obviously solve the function optimization and engineering optimization problems and has the merit of balancing the exploration and exploitation in the optimization process and enhances the convergence accuracy.
- Network intrusion detection based on IE-DBN model54
A deep belief network model based on information entropy (IE-DBN model) is proposed for network intrusion detection that obtained a higher detection accuracy and a lower false alarm rate than the conventional back propagation neural network and DBN network model.
- 54
- Learnability Matters: Active Learning for Video Captioning47
A novel active learning algorithm that takes three complementary aspects, namely learnability, diversity, and uncertainty, into account, and is capable of leveraging knowledge from humans in a more effective yet intellectual manner is designed.
- Investigating Examination Behavior of Image Search Users46
An eye-tracking study to investigate users' examination behavior in image searches focuses on the impacts of factors in examination including position, visual saliency, edge density, the existence of textual information, and human faces in result images and shows that combining position and visual content features can improve prediction in image search.
- Breast Mass Detection in Digital Mammogram Based on Gestalt Psychology45
A novel framework is proposed to detect breast masses in digitized mammograms that can be seen as an effective trade-off between bottom-up sensation and top-down recognition methods.
- THUIR at TREC 2005: Enterprise Track.45
A novel method which is called document reorganization to solve the problem of locating expert for certain query topics by collecting and combines related information from different media formats to organize a document which describes an expert candidate.
- Improved Algorithm for Insulator and Its Defect Detection Based on YOLOX44
The shortcomings of IoU and the sensitivity of small targets to the model regression accuracy are concerns and an improved SIoU loss function was proposed based on the regular influence of regression direction on the accuracy.
- Leveraging Passage-level Cumulative Gain for Document Ranking41
This paper proposes the context-aware Passage-level Cumulative Gain (PCG), which aggregates relevance scores of passages and avoids the need to formally split a document into independent passages and contributes to improving ranking performance and providing more explainability for document ranking.
- A new dataset of dog breed images and a benchmark for finegrained classification41
An image dataset for fine-grained classification of dog breeds: the Tsinghua Dogs Dataset, which has only one dog in each image and provides annotated bounding boxes for the whole body and head is introduced.
- Feature-based method for document alignment in comparable news corpora40
Experimental results on the English-Chinese and English-Malay comparable news corpora show that the proposed Discrete Fourier Transform-based term frequency distribution feature is very effective and contributes 4.1% and 8% to performance improvement over Pearson's correlation method on the two comparable corpora.
- A Label Dependence-Aware Sequence Generation Model for Multi-Level Implicit Discourse Relation Recognition38
This paper considers multi-level IDRR as a conditional label sequence generation task and proposes a Label Dependence-aware Sequence Generation Model (LDSGM) for it, which designs a label attentive encoder to learn the global representation of an input instance and its level-specific contexts.
- Aspect Sentiment Classification Towards Question-Answering with Reinforced Bidirectional Attention Network38
A Reinforced Bidirectional Attention Network (RBAN) approach is proposed to address two inherent challenges in ASC-QA, i.e., semantic matching between question and answer, and data noise.
- Adversarial Learning for Chinese NER From Crowd Annotations38
This paper proposes an approach to performing crowd annotation learning for Chinese Named Entity Recognition (NER) to make full use of the noisy sequence labels from multiple annotators, Inspired by adversarial learning.
- Yb3+/Tm3+/Ho3+ tri-doped YPO4 submicroplates: A promising optical thermometer operating in the first biological window38
Interestingly, the as-designed UC thermometer exhibits high temperature sensitivity (∼8.85% K −1 ) and good signal discriminability, which may be expected to find promising potential application for accurately detecting temperature of biological tissues.
- A prior feature SVM-MRF based method for mouse brain segmentation35
An automated method, called prior feature Support Vector Machine- Markov Random Field (pSVMRF), to segment three-dimensional mouse brain Magnetic Resonance Microscopy (MRM) images, which reduces training and testing time for SVM, while boosting segmentation performance.
- 35
- Open-ended Long Text Generation via Masked Language Modeling32
Two simple yet effective strategies for the iterative NAR model: dynamic sliding window attention (DSWA) and linear temperature decay (LTD) can alleviate long-distance collapse problems and achieve longer text generation with a flexible trade-off between performance and inference speedup.
- Semi-Supervised Bidirectional Long Short-Term Memory and Conditional Random Fields Model for Named-Entity Recognition Using Embeddings from Language Models Representations32
A semi-supervised deep learning model named SCRNER (Semi- supervised model for Cultural Relics’ Named Entity Recognition) that utilizes the bidirectional long short-term memory (BiLSTM) and conditional random fields (CRF) model trained by seldom labeled data and abundant unlabeled data to attain an effective performance.
- The nature of novelty detection32
The computational nature of the task currently overlooked by the Novelty community—Novelty as a combination of partial overlap (PO) and complete overlap (CO) relations between sentences is revealed.
- A Hybrid Multimodal Data Fusion-Based Method for Identifying Gambling Websites31
The experimental results indicate that OCR text on the webpage screenshots has strong semantic features and the proposed hybrid multimodal data fusion based method can effectively improve the performance in identifying gambling websites, with accuracy, precision, recall, and F1-score all over 99%.
- An LSTM-CNN attention approach for aspect-level sentiment classification31
An Aspect-based LSTM-CNN Attention model that combines LSTM with CNN for simultaneously leveraging LSTM to handle long-range dependencies and CNN’s ability to identify local features to improve the accuracy of aspect-level sentiment classification.
- Using active learning to expand training data for implicit discourse relation recognition31
Experimental results show that expanding the training set with small-scale carefully-selected external data yields substantial performance gain, with the improvements of about 4% for accuracy and 3.6% for F-score, which allows a weak classifier to achieve a comparable performance against the state-of-the-art systems.
- 30
- A Survey on Human Preference Learning for Large Language Models28
This survey reviews the progress in exploring human preference learning for LLMs from a preference-centered perspective, covering the sources and formats of preference feedback, the modeling and usage of preference signals, as well as the evaluation of the aligned LLMs.
- Incorporating Retrieval Information into the Truncation of Ranking Lists for Better Legal Search28
A ranking list truncation model for legal case retrieval, LeCut, which utilizes contextual features of the retrieval task to capture the semantic-level similarity between documents and decides the best cut-off position with attention mechanisms and a Joint Optimization of Truncation and Reranking (JOTR) framework based on LeCut to improve the performance of truncation and retrieval tasks simultaneously.
- Investigating User Behavior in Legal Case Retrieval28
A better understanding of user interactions in the legal case retrieval process is provided, which can benefit the design of the corresponding retrieval systems to support legal practitioners.
- Sentiment Classification towards Question-Answering with Hierarchical Matching Network28
A high-quality annotated corpus with specially-designed annotation guidelines for QA-style sentiment classification is created and a three-stage hierarchical matching network is proposed to explore deep sentiment information in a QA text pair.
- Tabu search algorithm for RP selection in PIM-SM multicast routing28
This paper proposes a multicast RP selection algorithm based on tabu search in PIM-SM multicast routing by considering both cost and delay, which shows that good performance is achieved in multicast cost, end-to-end delay and register delay.
- Learning to Reweight for Generalizable Graph Neural Network27
This paper proposes the Learning to Reweight for Generalizable Graph Neural Network (L2R-GNN) to enhance the generalization ability of GNNs for achieving satisfactory performance on unseen testing graphs that have different distributions with training graphs.
- A Unified Span-Based Approach for Opinion Mining with Syntactic Constituents27
The experimental results show that the proposed unified span-based approach achieves significant improvements over previous works in the exact F1 score and reduces the number of wrongly-predicted opinion expressions and roles, showing the effectiveness of the method.
- 26
- Improvement of extracellular lipase production by a newly isolated Yarrowia lipolytica mutant and its application in the biosynthesis of L-ascorbyl palmitate26
Under the optimal cultivation conditions, the scaled-up production of lipases were performed, and the lipase activity of Y. lipolytica Mut-96 was enhanced nearly 5.5-fold compared with that of Y., and the characteristics of the purifiedlipases were studied.
- 25
- Towards Accurate and Consistent Evaluation: A Dataset for Distantly-Supervised Relation Extraction24
A new dataset NYTH is built, where the DS-generated data is used as training data and annotators are hired to label test data and the experimental results show that human-annotated data is necessary for evaluation of distantly-supervised relation extraction.
- Human-Like Decision Making: Document-level Aspect Sentiment Classification via Hierarchical Reinforcement Learning24
This paper proposes a new Hierarchical Reinforcement Learning (HRL) approach to DASC that incorporates clause selection and word selection strategies to tackle the data noise problem in the task of DASC.
- Lexicon-Based Sentiment Analysis on Topical Chinese Microblog Messages24
A lexicon-based sentiment analysis method is introduced and a Weibo Lexicon with representative topical words and out-of-vocabulary (OOV) words, which are usually informal and not existing in formal dictionaries are constructed.
- THUIR at TREC 2003: Novelty, Robust and Web *23
A new IR system named TMiner has been built on which all experiments have been performed and a proposed Primary Feature Model (PFM) has been proposed and combined with BM2500 term weighting, which led to encouraging results.
- 22
- Axiomatically Regularized Pre-training for Ad hoc Search22
This work provides insights into improving the interpretability of pre-trained models and the guidance of incorporating IR axioms or human heuristics into pre-training methods.
- TREC-10 Web Track Experiments at MSRA.21
In TREC-10, Microsoft Research Asia (MSRA) participated in the Web track (ad hoc retrieval task and homepage finding task) and developed a method called collection refinement for QE, which investigated the use of two kinds of link information, link anchor and link structure.
- SelfMix: Robust Learning Against Textual Label Noise with Self-Mixup Training20
This work presents SelfMix, a simple yet effective method, to handle label noise in text classification tasks, which utilizes the dropout mechanism on a single model to reduce the confirmation bias in self-training and introduces a textual level mixup training strategy.
- GATSum: Graph-Based Topic-Aware Abstract Text Summarization20
The result shows that GTASum is superior to many extractive and abstract approaches in terms of ROUGE measurement and proves that the model has the ability to capture the original subject and the correct information and improve the factual accuracy of the summarization.
- 20
- 20
- 20
- 19
- Training Deep Ranking Model with Weak Relevance Labels19
The experimental results indicate that the ranking models trained with weak relevance labels are able to utilize large scale of behavior data and they can get similar performance compared to the ranking model trained based on relevance labels from external assessors, which are supposed to be more accurate.
- achieving efficient dynamic cryptographic access control in cloud storage19
Experiments show that HCRE can handle dynamic access policies in a more e-ficient way and does not reveal extra information of the plaintext to the CSP,thus it does no harm to the data confidentiality.
- Unleashing LLM Reasoning Capability via Scalable Question Synthesis from Scratch18
ScaleQuest introduces a two-stage question-tuning process comprising Question Fine-Tuning (QFT) and Question Preference Optimization (QPO) to unlock the question generation capabilities of problem-solving models and provides the open-source community with a practical solution to enhance the mathematical reasoning abilities of LLMs.
- 18
- 18
- 18
- 17
- Early Exit with Disentangled Representation and Equiangular Tight Frame17
An adapter method is proposed to decouple the two distinct types of representation and further introduces a non-parametric simplex equiangu-lar tight frame classifier (ETF) for improvement.
- 17
- 16
- A Cooperative Neural Information Retrieval Pipeline with Knowledge Enhanced Automatic Query Reformulation16
This paper presents a neural information retrieval pipeline that integrates cooperative learning of query reformulation and neural retrieval models and simultaneously optimize the quality of reformulated queries and ranking performance with an alternate training strategy where query reformulator and neural retrieved model learn from the feedback of each other.
- RG-NBEO: a ReliefF guided novel binary equilibrium optimizer with opposition-based S-shaped and V-shaped transfer functions for feature selection16
A ReliefF-guided novel binary equilibrium optimizer (RG-NBEO) is proposed for feature selection and can effectively improve the classification accuracy in most cases.
- LLMs Meet Long Video: Advancing Long Video Question Answering with An Interactive Visual Adapter in LLMs15
The proposed video-LLM facilitates a comprehensive understanding of long video content through appropriate long video modeling and precise visual interactions, and shows that the interactive visual adapter significantly improves the performance of video LLMs on long video QA tasks.
- 14
- 14
- 14
- User Browsing Graph: Structure, Evolution and Application.14
This paper tries to give a quantitative analysis on the difference in graph structure between hyperlink graph and user browsing graph, and then finds out why link analysis algorithms perform better on the browsing graph.
- FSAR: Federated Skeleton-based Action Recognition with Adaptive Topology Structure and Knowledge Distillation13
This paper pioneers a novel Federated Skeleton-based Action Recognition (FSAR) paradigm, which enables the construction of a globally generalized model without accessing local sensitive data, and introduces an Adaptive Topology Structure (ATS), separating generalization and personalization by learning a domain-invariant topology shared across clients and a domain-specific topology decoupled from global model aggregation.
- A Novel Approach for Detecting Rotational Angles of a Precision Spherical Joint Based on a Capacitive Sensor13
Simulation results show that the variation in the simulated capacitance values of the capacitors is similar to that of the theoretical values, suggesting the feasibility and effectiveness of the proposed capacitive detection method for rotational angles of spherical joints.
- 13
- 12
- 12
- 12
- 12
- Driving Risk Assessment in Work Zones Using Cloud Model12
The model proposed in this paper can accurately assess the driving risks in work zones using a more scientific and intuitive approach, which provides an excellent tool to design safe expressway work zones.
- 12
- 12
- 12
- Hierarchical learning strategy in semantic relation extraction12
Two classifier learning approaches, i.e. the simple perceptron algorithm and the state-of-the-art Support Vector Machines, are applied using the hierarchical learning strategy to deal with the data sparseness problem in semantic relation extraction.
- Improved Feature Selection and Redundance Computing - THUIR at TREC 2004 Novelty Track.12
Research on this year’s novelty track mainly focused on text feature selection and reduction, improved sentence classification in finding relevant information, efficient sentence redundancy computing and effective result filtering.
- THU TREC2002 Web Track Experiments12
A genetic algorithm based dynamic parameter learning approach is proposed and implemented for key resource selection based on the assumption that a key resource is more likely to link to multiple relevant documents.
- Improving primary healthcare quality in China through training needs analysis11
The diagnosis and treatment capability levels of healthcare professionals directly affect the overall quality of medical services and special funding support and training in the latest medical guidelines, new technologies/skills, and advanced management concepts are important to improve the composition of titles, education, and professional technicians.
- 11
- A Novel Three-Factor Authentication Scheme with High Security for Multi-Server Environments11
This study presents an improved biometrics-based three-factor authentication scheme for multi-server environments that inherits most of the advantages of the original scheme and introduces digital signature to address the common security problem.
- 11
- A Hybrid Framework for Session Context Modeling11
A novel Hybrid framework for Session Context Modeling (HSCM), which realizes session-level multi-task learning based on the self-attention mechanism and not only outperforms strong baseline solutions such as HiNT, CARS, and BERTserini in document ranking, but also performs significantly better than most existing query suggestion methods.
- Two-branch encoding and iterative attention decoding network for semantic segmentation11
This paper proposes a two-branch encoding and iterative attention decoding semantic segmentation model for deep convolutional neural networks and proposes a channel position attention module and a boundary residual attention module to learn different position and boundary features, which can enrich the target boundary position information.
- 11
- Improving Relation Extraction with Relational Paraphrase Sentences11
This work proposes a novel model to learn the information of diverse relation expressions via enriching diverse expressions by relational paraphrase sentences via a joint learning framework and shows that this approach is effective to improve the performance on relation extraction, even compared with a strong baseline.
- 11
- The first experimental demonstration of a DREAM-based large-scale optical transport network with 1000 control plane nodes11
Experimental results show the good performance of DREAM, a scalable framework named Dual Routing Engine Architecture in Multi-layer, multi-domain and multi-constraints scenarios for the routing issue in large-scale dynamic optical networks.
- 10
- A Read-and-Select Framework for Zero-shot Entity Linking10
A read-and-select (ReS) framework by modeling the main components of entity disambiguation, i.e., mention-entity matching and cross-entity comparison, achieves the state-of-the-art performance on the established zero-shot EL dataset ZESHEL with a 2.55% micro-average accuracy gain.
- 10
- 10
- Text GCN-SW-KNN: a novel collaborative training multi-label classification method for WMS application themes by considering geographic semantics10
A novel multi-label text classification method, Text GCN-SW-KNN, based on geographic semantics and collaborative training to improve classification accuracy and can be extended to other multi- label text classification scenario for better supporting metadata enhancement and geospatial resource discovery in Earth Science domain.
- Attention-based encoder-decoder networks for workflow recognition10
A novel workflow recognition framework to recognize the behavior of working subjects based on the well-designed encoder-decoder structure, termed as AWR, is proposed and the effectiveness of the workflow recognition system is validated in a real-world production scenario.
- An Information Hiding Algorithm for HEVC Videos Based on PU Partitioning Modes10
The proposed algorithm can be combined with existing DCT-based or intra-prediction-based algorithms to greatly increase the embedding capacity while keeping the quality of the video sequence not being affected.
- Using Entity Relation to Improve Event Detection via Attention Mechanism10
This paper proposes a novel event extraction method to exploit relation information for event detection (ED), due to the potential relevance between entity relation and event type, and systematically investigates the effect of relation representation between entities.
- CLGVSM: Adapting Generalized Vector Space Model to Cross-lingual Document Clustering10
Experimental results on benchmarking data set show that the proposed CLGVSM is very effective for cross-document clustering, outperforming the two strong baselines vector space model (VSM) and latent semantic analysis (LSA) significantly; and the new feature selection method can further improve CLGV SM.
- DF or IDF? On the Use of Primary Feature Model for Web Information Retrieval10
A new PFS query term weighting scheme is proposed, which takes document frequency (DF) into account instead of the traditional IDF factor and improves the system performance consistently and effectively in terms of recall, top n precision and mean average precision.
- Clustering Concepts10
RefViz 1.0 allows researchers and students to visualize and explore thematic and conceptual connections among references from scientific literature.
- Development of high-efficiency superparamagnetic drug delivery system with MPI imaging capability9
This study contributes to the development of local, rapid, and precise drug delivery and the visualization of this process during cancer therapy, and further research on MPI/delivery synchronization technology is planned for the future.
- 9
- On Annotation Methodologies for Image Search Evaluation9
This work is the first to systematically study how diverse factors in data annotation impact image search evaluation, and suggests different strategies for exploiting the crowdsourcing to get data annotated under different conditions.
- A Novel Method for the Micro-Clearance Measurement of a Precision Spherical Joint Based on a Spherical Differential Capacitive Sensor9
Simulation results indicated that the proposed method based on SDCS was feasible and effective for the micro-clearance measurement of the precision spherical joints with small eccentricity.
- 9
- A Clustering Approach for Privacy-Preserving in Social Networks9
This paper introduces a novel approach for privacy-preserving considering both attribute and structure information, and defines the entropy to measure capability of preserving sensitive attributes.
- Measurement of Trust Transitivity in Trustworthy Networks9
This paper abstracts the trust network as a weighted digraph, and proposes a matrix-based method to compute the max-mean trust degree, which can be done polynomially, but may produce non-simple paths.
- Optimising inter-view prediction structure for multiview video coding with minimum spanning tree9
A new systematic method to construct a near-optimal inter-view prediction structure for multiview video coding is proposed based on the directed minimum spanning tree of inter-View prediction cost graph.
- Incremental Learning for Profile Training in Adaptive Document Filtering.9
An incremental learning method which selects pseudo positive documents in less bias from a few initial positive training documents is developed, which can be helpful for profile training, while the new language model perform not well.
- Test-time Adaptation for Machine Translation Evaluation by Uncertainty Minimization8
This paper aims to address the inference bias of neural metrics through uncertainty minimization during test time, without requiring additional data, by employing the prediction uncertainty of the current data as a signal to update a small fraction of parameters during testTime.
- DMPNet: densely connected multi-scale pyramid networks for crowd counting8
A Densely Connected Multi-scale Pyramid Network (DMPNet) for count estimation and the generation of high-quality density maps, which can extract multi-scale features of the crowd effectively while keeping the resolution of the input feature map and the number of channels unchanged.
- Action detection with two-stream enhanced detector8
This paper devise an appearance and motion detectors in two-stream manner to detect actions, which take k consecutive RGB frames and optical flow images as input respectively, and introduces anchor refinement sub-module with feature alignment to improve the feature presentation capabilities.
- Investigating human reading behavior during sentiment judgment8
A lab-based user study in which users’ fine-grained reading behaviors during microblog sentiment classification are recorded with an eye-track device and it is found that sentiment judgment is more like an auxiliary task of content comprehension for humans.
- Semantic Role Labeling with Heterogeneous Syntactic Knowledge8
This work proposes to encode heterogeneous syntactic knowledge for SRL from both explicit and implicit representations, and introduces graph convolutional networks to explicitly encode multiple heterogeneous dependency parse trees.
- Emotion Detection with Neural Personal Discrimination8
Experimental results show the usefulness of personal attributes, and the effectiveness of the proposed NPD approach in capturing such personal attributes with significant gains over the state-of-the-art models.
- EARS 20198
This workshop focuses on the research and application of explainable recommendation, search, and a broader scope of IR tasks and will generate insightful debates about the recent regulations regarding AI interpretability to a broader community including but not limited to IR, machine learning, AI, Data Science, and beyond.
- 8
- 8
- A Unified Framework for Emotional Elements Extraction Based on Finite State Matching Machine8
This work designed a unified framework based on finite state matching machine to deal with the problems of emotional element extraction and achieves high accuracy and scalability in emotional elements extraction.
- 8
- Research on Network Virtualization and Virtual Network Mapping Algorithm8
A new mapping algorithm PBMC is proposed, which is based on path based multi-commodity problem model to resolve the problem of hop count limit and network resource distribution balance, and simulation results indicate that PBMC performs better in virtual network request acceptance ratio and resource utilization ratio.
- 8
- Citation Managers and Citing-Cited Data.8
This article evaluates how two citation managers, EndNote and RefWorks, provide ways to store, display, and retrieve cited and citing reference lists, and suggests enhancements with respect to cited and cite reference data that should be adopted by any citation management software that currently lacks them.
- A Multimodal In-Context Tuning Approach for E-Commerce Product Description Generation7
A simple and effective Multimodal In-Context Tuning approach, named ModICT, which introduces a similar product sample as the reference and utilizes the in-context learning capability of language models to produce the description.
- 7
- RACMF: robust attention convolutional matrix factorization for rating prediction7
The experimental results show that the RACMF model has good prediction ability, even when the rating data are sparse or the scale of rating data is large, and compared with the state-of-the-art model PHD, the present model has increased the accuracy rate.
- 7
- A Virtual Network Mapping Algorithm Based on Time7
A time-based VN embedding algorithm is proposed and a probability model is formulated to obtain the maximum probability that the available resources of substrate network can be used by succeeding VN requests.
- 7
- 7
- 7
- UniMoE-Audio: Unified Speech and Music Generation with Dynamic-Capacity MoE6
UniMoE-Audio, a unified speech and music generation model within a novel Dynamic-Capacity Mixture-of-Experts (MoE) framework, achieves state-of-the-art performance on major speech and music generation benchmarks, but also demonstrates superior synergistic learning, mitigating the performance degradation typically seen in naive joint training.
- Improving Rationality in the Reasoning Process of Language Models through Self-playing Game6
This paper designs a Critic-Discernment Game (CDG) in which a prover first provides a solution to a given problem and is subsequently challenged by critiques of its solution, which demonstrates that CDG training can significantly improve the ability of well-aligned LLMs to comprehend their reasoning process.
- 6
- Automatic classification of heart failure based on Cine-CMR images6
The proposed 4D-SSLHF can mine 3D spatial information and temporal information in Cine-CMR images well and accurately classify different categories of HF, showing its potential to assist physicians in choosing personalized treatment.
- Dense-scale dynamic network with filter-varying atrous convolution for semantic segmentation6
A filter-varying atrous convolution (FAC) to efficiently enlarge the per-pixel receptive fields pertaining to various objects and a dense-scale dynamic network (DsDNet) that simultaneously enhances the localization accuracy and reduces the effect of large-scale variations of the object, by assigning FAC to different spatial locations at dense scales.
- [Retracted] Application of Artificial Intelligence Combined with Three‐Dimensional Digital Technology in the Design of Complex Works of Art6
The Surface Wavefront Reconstruction on Fast Fourier Transform Method is implemented, evaluating the performance of a combination of AI and three-dimensional technology, and it is observed that the proposed SWRFFT method provides a 2.47% increased accuracy rate in the correlation of the image with the frequency generation.
- 6
- Fast Multi-keywords Search over Encrypted Cloud Data6
This paper design and implement an SE scheme that allows the MKS indices to be pre-built secretly at the server side, making fast MKS on very large databases possible.
- THUSAM at NTCIR-11 IMine Task6
The approaches and results in NTCIR11 IMine task are described, which include rerank the result lists with HITS algorithm and then adopt a pruning exhaustive search algorithm to generate diversied result lists.
- 6
- 5
- Graph embedded low‐light image enhancement transformer based on federated learning for Internet of Vehicle under tunnel environment5
This paper deploys an image enhancement module at the terminal of the IoV to alleviate the low illumination influence and confirms that the proposed graph embedded Transformer model can effectively enhance the detail information of the low‐light image, which can greatly improve the following tasks in IoV.
- Deep Convolutional Neural Network Compression Method: Tensor Ring Decomposition with Variational Bayesian Approach5
A TR network compression method by Variational Bayesian (TR-VB) is proposed based on the Global Analytic Solution of Empirical Variational Bayesian Matrix Factorization (GAS of EVBMF), which gives the best results in terms of Top-1 accuracy, parameters, and training time under different compression levels.
- 5
- 5
- 5
- Cognition-driven multimodal personality classification5
To tackle the inherent noise challenges in this CMPC task, this work proposes a tailored reinforcement learning approach, namely, multi-agent SelectNet, aiming to integrate the opinion-word and image-region selection strategies to select informative opinion- word andimage-region features for CMPC.
- A novel video target tracking method based on lie group manifold5
This paper integrates the particle filtering algorithm with the Lie group theory into an innovative video target tracking algorithm that describes the target features with covariance matrix, and expresses the position and shape changes of the target, considering the spatial and temporal statistics.
- Hop-Limit Path Mapping Algorithm for Virtual Network Embedding5
A fast approximation path mapping algorithm to address the problem of path mapping with hop limit by formulating such virtual link mapping problem as a path-flow mathematical programming model, which aims to minimize the maximum link load factor.
- 5
- 5
- 5
- On optimal binary signed digit representations of integers5
What the exact number of optimal BSD-R of an integer is and how to generate them entirely is studied and it is shown which kinds of integers have the maximum number.
- 5
- 4
- 4
- Make Imagination Clearer! Stable Diffusion-based Visual Imagination for Multimodal Machine Translation4
A stable diffusion-based imagination network integrated into a multimodal large language model (MLLM) to explicitly generate an image for each source sentence, thereby advancing multimodal MT.
- 4
- VideoVista-CulturalLingo: 360$^\circ$ Horizons-Bridging Cultures, Languages, and Domains in Video Comprehension4
VideoVista-CulturalLingo is the first video evaluation benchmark designed to bridge cultural, linguistic, and domain divide in video comprehension, and observes that existing models perform worse on Chinese-centric questions than Western-centric ones, particularly those related to Chinese history.
- 4
- Dynamic bandwidth allocation algorithm based on extreme learning machine in passive optical networks4
A machine learning-based DBA algorithm that utilizes extreme learning machine (ELM) to predict the future load conditions of ONUs and achieves low delay of 125 μs, meeting the 3GPP requirements for delay below 250 μs, and the jitter remains below 80 μs consistently.
- Human pose estimation based on parallel atrous convolution and body structure constraints4
A parallel atrous convolutional network with body structure constraints (PAC-BCNet) is designed to address the problem of human pose estimation and is evaluated on two widely usedhuman pose estimation benchmarks (MPII and LSP).
- Proposal-Based Graph Attention Networks for Workflow Detection4
A generic proposal based Graph Attention Networks for workflow detection is proposed, an efficient and effective action proposal method is firstly employed to generate workflow proposals, and within-category and between-category attention are incorporated to learn long-range and dynamic dependencies respectively.
- 4
- 4
- Experimental demonstration of software-defined optical network for heterogeneous packet and optical networks4
This work proposes a control architecture for packet and optical network convergence, including how to be compatible with the existing optical network control architecture, and demonstrates a series of network performance parameters under various network environments.
- 4
- A differential evolution-based memetic algorithm for project scheduling problems4
Simulation results on real-world PSP instances validate the superiority of DEMA over state-of-the-art search-based approaches and the effectiveness of new mechanisms designed in DEMA.
- 4
- A general collaborative filtering framework based on matrix bordered block diagonal forms4
A novel CF framework based on Bordered Block Diagonal Form (BBDF) matrices attempting to meet the challenges of data sparsity and scalability is presented, which helps to improve prediction accuracy in collaborative filtering tasks.
- Hierarchical virtual network mapping algorithm for large-scale network virtualisation4
The proposed hierarchical algorithm could find the best trade-off between the number of optimisation iterations (computation time) and cost by selecting an appropriate number of iterations and shows that nearly optimal performance as well as improved VN mapping request acceptance ratio and revenue can be achieved.
- 4
- THUIR at TREC 2004: Genomics Track *4
This is the first time that THUIR participates in TREC Genomics Track and took part in both Ad hoc retrieval task and Categorization task, focusing on category of organism retrieval strategy and primary feature model.
- 3
- 3
- 3
- Temporal Action Localization With Coarse-to-Fine Network3
Compared with previous state-of-the-art techniques, the proposed coarse-to-fine network can asymptotically approach fine-grained action boundary and demonstrates outstanding improvements of the method when compared with the prior methods on various video action parsing tasks.
- 3
- Leveraging Document-Level and Query-Level Passage Cumulative Gain for Document Ranking3
This paper investigates how information gain accumulates with passages and proposes a BERT-based sequential model called Passage-level Cumulative Gain Model (PCGM), which outperforms most existing ranking models and contributes to improving ranking performance and providing more explainability for document ranking.
- Multiway dynamic mask attention networks for natural language inference3
The Multiway Dynamic Mask attention Network (MDMAN) is proposed, which has two modules: a dynamic mask selector and a multi-attention encoder that achieves significant improvement over DSAN.
- A new three-factor authentication scheme overcome repeat registered attack for wireless sensor networks3
Thorough analysis shows that the presented scheme can provide stronger security and slight lower computation at client than Das’s protocol and can overcome repeat registered attack and achieve sensor nodes’ identity anonymity protection.
- A classification algorithm based on weighted ML-kNN for multi-label data3
The main idea is to assign different weights to each label according to the proportion of labels and mutual information of the spatial distribution of unseen instances to training instances, which can reduce the probability of misjudgement of the unseen instance's label set.
- 3
- 3
- Semi-Supervised Dependency Parsing3
This chapter discusses the development of semi-supervised Dependency Parsing Approaches, and some of the techniques used to develop these models.
- 3
- A Comparison of Five HSV Color Selection Interfaces for Mobile Painting Search3
Which color palette(s) enable users to complete task faster, how good these resulting drawings were for searching, and what were user experiences are investigated.
- Performance Analysis of Cognitive Radio Networks Adopting Packet Discarding or Buffering Policy3
An analytical model is developed to derive the formulas for both calllevel performance measures and packet level performance measures for SUs and shows that theoretical models are consistent with simulation results.
- 3
- Grammatical Phrase-Level Opinion Target Extraction on Chinese Microblog Messages3
This work proposes a unified approach that concentrates on phrase-level target extraction of opinion target extraction on Chinese microblog messages by classifying grammatical relations with a cost-sensitive classifier that enhances performance of unbalanced data and prune the noisy targets by empirically summarized rules.
- A Speech Recognition System Based Improved Algorithm of Dual-template HMM3
Experimental result shown the speech recognition rate of large vocabulary of non-specific people was greatly improved and the model and recognition algorithm of HMM got be improved based on the traditional the HMM.
- 3
- Name Origin Recognition Using Maximum Entropy Model and Diverse Features3
This paper casts the name origin recognition as a multi-class classification problem and approach the problem using Maximum Entropy method, and investigates the use of different features, including phonetic rules, ngram statistics and character position information forName origin recognition.
- 3
- GCS-YOLO11: a lightweight YOLO model based on improved MSR image enhancement and GhostConv module for traffic sign recognition2
Before model detection, preprocesses the low-quality images in the dataset to enhance the features of the target to be detected in the images and then inputs them into the improved YOLO model for traffic sign recognition to improve the accuracy and efficiency of model detection.
- 2
- Enhancing Visual-Language Prompt Tuning Through Sparse Knowledge-Guided Context Optimization2
The innovative concept of SparseKnowledge-guided Context Optimization (Sparse-KgCoOp), which seeks to narrow the gap between the textual embeddings produced by both the dynamic prompts and the manually devised ones, thus preserving the foundational knowledge while maintaining adaptability, is proposed.
- 2
- 2
- The Ladder in Chaos: A Simple and Effective Improvement to General DRL Algorithms by Policy Path Trimming and Boosting2
This paper studies how the policy networks of typical DRL agents evolve during the learning process by empirically investigating several kinds of temporal change for each policy parameter, and proposes a simple and effective method, called Policy Path Trimming and Boosting (PPTB), as a general plug-in improvement to DRL algorithms.
- Test-Time Adaptation with Perturbation Consistency Learning2
PCL is proposed, a simple test-time adaptation method to promote the model to make stable predictions for samples with distribution shifts, and can achieve higher or comparable performance with less inference time over strong PLM backbones and previous state-of-the-art TTA methods.
- Occlusion-robust workflow recognition with context-aware compositional ConvNet2
The proposed context-aware CA-CompNet can robustly detect workflow instance under occlusion in industrial environment, increasing the detection performance on MS COCO dataset by 4.6% in absolute performance compared to the advanced CenterNet.
- 2
- 2
- 2
- SogouQ: The First Large-Scale Test Collection with Click Streams Used in a Shared-Task Evaluation2
This chapter introduces a real Chinese query log dataset, SogouQ, which was released in 2010 for the NTCIR-9 Intent task and has a broader impact on other research areas, such as natural language processing and social science.
- Cascaded Multi-Channel Feature Fusion for Object Detection2
Experimental evaluation of the proposed CM-FPN approach has shown that the proposed method produces higher object detection accuracy.
- An improved neural network for domain adaptive Chinese word segmentation2
An improved neural-based CWS model is proposed, which incorporates the common lexicon and unlabeled data into BERT and is very strong in domain adaptability.
- Recommendation algorithm based on user attributes and tag preferences2
A UT-CF algorithm that combines user attributes and tag preferences is proposed, and the K-means clustering algorithm based on the principle of maximum distance is introduced to reduce the search space of similar users.
- 2
- 2
- 2
- 2
- Hierarchical Structure Information and Its Expression for Image Features Extraction and Processing of GRIN Lens End2
H hierarchical structure and its expressional method of image information are described based on the image processing experiments of gradient-index (GRIN) lens end face, and innovative experiment methods and promising results are presented.
- 2
- A Recognition Method of Pedestrians’ Running in the Red Light Based on Image2
The method of adaptive background model of mixture Gaussian is used to extract the moving objects in the video and identify whether the pedestrians are running in the red lights or not, according to the transportation signals and allocated walking areas.
- Method for personalized user profiling in social tagging systems2
A new user modeling method was proposed, which introduced co-occurrence techniques into the naive approach and added some tag pairs on the basis of naive approach to user model, which reflects tags relationship and considers tags weight.
- 2
- Secure Range Query with Single Assertion on Encrypted Data2
A schema of range query on encrypted data is proposed, which reduces the amount of information leakage by only asserting the interval index once, and is multiplied with invertible matrix to guarantee the security issues of the sensitive data and the search range.
- 2
- 2
- 2
- 2
- 1
- 1
- Network pharmacology combined with molecular docking and experimental validation reveals the anti-pharyngitis mechanism of Humei Qingyan granule1
HMQY exerts its therapeutic effects on pharyngitis primarily by inhibiting the PI3K-Akt/NF-κB signaling pathway, and preliminarily validated the active ingredients, key targets, and potential pathways of HMQY in the treatment of pharyngitis.
- Multi-granularity transformer fusion for temporal action localization1
The MGTF builds a multi-granularity feature fusion pipeline based on transformer, and uses a direct set prediction strategy to generate action instances to reduce computation cost and to reduce computation cost, design temporal shift attention to adaptively focus on a sparse set of key segments.
- 1
- 1
- 1
- 1
- Protocol for preparing patterned superlattice structures by printing assembly technology for multi-channel detection1
A protocol to prepare inks for constructing patterned superlattice structures using print assembly techniques and detail procedures for preparing a multichannel surface-enhanced Raman scattering sensor and evaluating its performance are presented.
- Remu-Net: Multi-Branch Net Framework for 3D Brain Tumor Segmentation1
An end-to-end, more efficient brain tumor MRI segmentation model, REMU-Net, for the problems of multi-scale feature extraction and difficulty in small target feature extraction in 3D brain tumor image segmentation is proposed.
- Multi-metric learning by a pair of twin-metric learning framework1
Under different evaluation criteria, experiments show that the proposed TMML is superior to the single metric learning methods, and achieves better performance than other state-of-the-art multi-metric learning methods in most cases.
- Neural Emotion Detection via Personal Attributes1
Experimental results show the usefulness of personal attributes in emotion detection, and the effectiveness of the proposed JPA and NPD approaches in capturing personal attributes over the state-of-the-art statistic and neural models.
- 1
- 1
- 1
- A Hybrid Model with Pre-trained Entity-Aware Transformer for Relation Extraction1
A novel hybrid model that combines Piece-wise Convolutional Neural Network (PCNN) and Entity-Aware Transformer to extract local features and learn the dependencies between distant positions jointly is proposed.
- A quasi-conformal mapping-based data augmentation technique for improving deep learning techniques on brain tumor segmentation1
This work introduces a novel differential geometry-based quasi conformal (QC) mapping augmentation technique to augment the brain tumor images and demonstrates the efficiency and efficacy of the proposed method.
- A vehicle detection algorithm in complex traffic scenes1
An improved Faster R-CNN algorithm for accurate vehicle detection using feature maps of depth images to supplement vehicle details by adding a depth channel into the detection model and evaluating the newly proposed approach using KITTI dataset.
- 1
- Transformer Based Memory Network for Sentiment Analysis of Chinese Weibo Texts1
A new Transformer (a new neural network architecture based on self-attention mechanism) based memory network (TF-MN), is introduced and the result of the experiment proves that the model reaches better accuracy than the state-of-the-art model.
- Research on encoding and decoding algorithms of non-binary LDPC code and FPGA implementation1
FFT-BP decoding algorithm of non-binary LDPC code is studied, and on this basis, an improved FFT-BP algorithm is proposed which can effectively reduce the average iteration number and encode and decode algorithm of non-binary LDPC code with FPGA is achieved.
- A De-anonymization Attack for Social Network Graph Based on Structural and Node Feature Similarity1
A novel de-anonymization method based on the structure of the social network graph of users and the characteristics of user nodes to disclose users’ private information is proposed, which outperforms in accuracy and execution time.
- A Novel Tracking-by-Detection Method with Local Binary Pattern and Kalman Filter1
The local binary pattern (LBP) is integrated with the cascade classifiers, and a new classifier named ULBP (Uniform Local Binary Pattern) classifiers are defined, which will work instead of the nearest neighbor classifier in TLD algorithm when the object appearance has rich texture features.
- 1
- Virtual Network Mapping Algorithm Based on Load Balancing1
A two-dimensional discrete weighted model based on time and resources, and a mathematical programming model of minimizing the degree of two- dimensional load balancing are proposed that can increase the acceptance ratio and the revenue by the substrate network in the long term.
- 1
- A Preliminary Examination of the User Behavior in Query-by-Drawing Portrait Painting Search on Mobile Devices1
A Query-by-Drawing (QbD) mobile application is developed, and a user study is conducted to explore the search behavior of painting search by drawing on the touchscreen phone to analyze the patterns of query formulation and query modification.
- 1
- Pornography Detection with the Wisdom of Crowds1
Inspired by the idea that different pornography Web pages may be oriented by similar search keywords, a label propagation method on click-through bipartite graph is proposed which can locate pornography web pages from a small set of manually labeled seed pages.
- 1
- Research for buffer overflow vulnerabilities based on multi-dimensional Fuzzing technology1
The results of testing Libpng validate that MFBOF is effective, and the paper gave its further improvement directions.
- Graph-theoretic algorithm for spectrum allocation in IEEE 802.22 wireless mesh network1
A new graph-theory algorithm is proposed that can improve the performance of spectrum allocation in IEEE 802.22 mesh networks and could be transformed into a weighted complete bipartite-graph by three processes.
- 1
- An Improved Scheme of the "Medium-Field" Multivariate Cryptosystem1
The idea of internal perturbation is imposed on the MFE cryptosystem to construct a new variant that can be used to cope with the second order linearization equation attack, and has some improvement in the aspect of the space of key and security compared with other existing schemes.
- News Page Discovery Policy for Instant Crawlers1
Empirical studies in a user access log for two months show that the proposed novel approach to discover news pages outperforms the traditional approach in both precision and recall.
- Identify Temporal Websites Based on User Behavior Analysis1
A large proportion of temporal pages are published by a small number of web sites providing news services, which should be crawled repeatedly with small intervals, and such temporal web sites of high freshness requirements can be identified by the algorithm based on user behavior analysis in page view data.
- 1
- 1
- Research on Product Image Design which Reflect Enterprise Image1
The design method based on the strong association of the products and corporate image was been looking for from analysis of the idea aspects and vision aspects of the Products.
- Learning-based Web Data Cleansing for Information Retrieval *1
A data cleansing algorithm is proposed by selecting an important type of high quality pages (key resources) on the web by selecting more than 98% of hyperlinks and covers about 90% of key information.
- 1
- Using Normal PC to Index and Retrieval Terabyte Document - THUIR at TREC 2004 Terabyte Track.1
An electronic ballast for a discharge lamp for restricting and stabilizing the lamps current and comprises an inductor, resonance capacitors and a capacitor coupled in parallel with the lamp, which is connected in series with the resonance circuit.
- An Efficient Multiple-Constraints QoS Routing Algorithm Based on Nonlinear Path Distance1
A QoS routing algorithm (called MIS-LB), which extends the definition of nonlinear path distances into the link-sharing and load balancing, and shows that it outperforms other multiple-constraints routing, such as TAMCRA and H_MCOP.
- 1
- Optimized TQ-MAP: An Adaptive Mapping Mechanism to Guarantee the Coherence of QoS Support from IP to OBS1
The proposed novel assembly mapping mechanism, called Optimized TQ-MAP, based on nonlinear programming and differential calculus, allocates the burst assembling capacity between classes fairly, efficiently and differentially, and matches IP QoS requirement with OBS QoS capacity as possible as it can.
- Collaborative Disassembly for Multiple Users in Virtual Reality–
This work presents Dynamic Parallel Disassembly and Tasking (DPDT), a novel framework tailored for immersive collaboration that achieves a computational speedup of up to 4.7 times while improving planning reliability by approximately 9.3 percentage points.
- –
- –
- Safety helmet wearing detection algorithm in complex scenarios based on improved YOLOv8n–
The Channel-Grouped Multi-Scale Convolution Module (CMSC) is integrated into C2f, using multi-scale convolution to enhance feature representation in foggy and dusty environments, reducing blurring.
- –
- –
- –
- DNN compression approach based on Bayesian optimization tensor ring decomposition–
A TR network compression method by Bayesian optimization (TR-BO) is proposed, which involves selecting rank via Bayesian optimization, compressing the neural network layer via TR decomposition using rank obtained in the previous step, and, finally, further fine-tuning the compressed model to overcome some of the performance loss due to compression.
- –
- –
- –
- –
- –
- –
- –
- –
- –
- A fixed single station bearing-only target motion parameter estimation method–
A weighted instrumental variable (WIV) motion parameter estimation method based on prior velocity information has strong robustness under large measurement angle error and velocity prior information error and its performance is theoretically analyzed.
- –
- A particle swarm algorithm combined with flower pollination algorithm–
The comparison of the optimization accuracy and the convergence speed on the four test functions proves the superiority of the flower pollination and particle swarm algorithm.
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- A neural joint model for rating prediction recommendation–
A neural joint model based on reviews, products category, and users’ co-purchase information for rating prediction recommendation called NMRP is proposed, which outperforms recently proposed baselines such as DeepCoNN, TransNets, and NARRE.
- –
- –
- –
- –
- –
- Image Processing Framework for Face Detection and Face Swapping in Group Photo Refinement–
This paper proposes a relatively reliable and complete framework to refine photos of human facial expressions of emotions among a set of group photos both partially and wholly based on existing techniques of face detection and face swapping.
- Relation extraction based on semantic dependency graph–
A BiLSTM-CNN-Attention model (BLCANN) based on semantic dependency graph to extract sentence features is proposed, which combines the structural and semantic features of the sentence, which contributes to distinguishing between positive and negative examples in multi-instance learning.
- –
- A Multi-keyword Searchable Encryption Based on B+ Tree–
A Multi-keyword Searchable encryption technology based on the B+ tree is proposed, which establishes indexes and retrieval trapdoor by using the space vector model and sort retrieved results according to the relevance score and the keyword matching.
- –
- Improved Method of Contention-Based Random Access in LTE System–
This work proposes a method on the basis of the existing alternatives to estimate the system load in advance to adjust the number of terminals and has an improvement on the existing power climbing strategy.
- –
- –
- Application of information technology in the model of nursing human resource allocation–
The nursing human resource allocation model built with information technology could export thehours of direct nursing in real time and configure nursing staff reasonably in line with the hours of nursing and can improve the quality of nursing, patients' satisfaction and nurses' job satisfaction.
- A Fast Prediction Algorithm for Sina Weibo Users with Time Correlation Cognition–
The rapid predicted classification algorithm based on correlation time cognition for weibo users in specific events is proposed, which can accomplish weibo user classification according to their behavior in the short time window, so as to establish a more accurate users behavior relation network.
- –
- An investigation of query-by-drawing image search on mobile devices–
A flexible platform is built that can be used for research in drawing-related fields, such as a new QbD technique or interface tests and psychological study, and a new way of categorising colour picker is proposed, which extends a scientific understanding of using current Query-by-Drawing techniques for real world image search.
- –
- Study on the Legal Status of Computer Network Crime and Its Legal Prevention–
Based on the analysis of computer network crime laws targeting on further exploration forComputer network crime law protection policy, some valuable recommendations for computer network security can be improved.
- –
- –
- Distributed Potential Field Based Routing with Priority Connection Probability for NDN–
This paper proposes a novel distributed PBR scheme with prioritized connection probability of the nodes calculated by the popularity and similarity, which will be used for the function of Fermi-Dirac distribution to calculate connection probability.
- –
- –
- –
- –
- –
- An integrated evaluation model of the safety controller's workload based on improved extension evaluation model–
An improved integrated evaluation model of the workload of an air traffic controller based on the matter-element is proposed to obtain the optimal scheme through the analysis of the relation of the various indexes of the safety control work by taking the terminal control area of Shanghai as a case study sample.
- –
- An improved cooperative spectrum detection algorithm for cognitive radio–
The paper proposes a method to optimize the weight on basis of the deflection coefficient criterion, and shows that the proposed algorithm has better performance even in low signal-to-noise ratio (SNR).
- Factors influencing knowledge contribution: An empirical investigation of social networking website users–
The algorithms for calculating similarity of Chinese sentences based on hybrid strategies, which integrate the question surface similarity, the question semantic similarity and the question-answer similarity based on latent semantic analysis(LSA) to find answers to a user's question.
- Image De-noising and Granularity Detection Based on Morphology–
Experimental results show that the image granularities, that are detected by using different structure elements, have great difference, and two-dimensional detection is better than one-dimensional direction detection.
- –
- Performance Analysis of Cooperative Spectrum Detection Based on Alamouti Coding–
Analysis results show that, as the quality of local user channel gets better, symbol Error Rate of reporting channel decreases with a coding gain more than 3 dB, and the minimum probability of false alarm decreases obviously.
- The Airport Runway Foreign Objects Detection Method Research Based on the Algorithm of SIFT–
Experimental results show that this method can not only handle the problems of tar-get losing evidently, which are induced by objects rotation and translation, but also has nice robustness to the conjunction of multi-targets in the process of object tracking.
- –
- –
- –
- Discussing on a WLAN wireless optimization strategy based on scene–
In the demand of eight WLAN hotspot scene, the paper refreshes the idea of the coverage, capacity, quality optimization, and gives targeted optimization measures and examples.
- An Energy-Balanced Routing Algorithm for ZigBee Audio Guide System in Ad Hoc Social Network–
The simulation results indicate that this energy-balanced routing algorithm can balance the entire network energy, extend the survival time of the whole network and increase the stability of the system relative to the original routing protocol.
- Enhancing the Flexibility of Kademlia under Churn–
Simulation results show that the routing table self-adaptive mechanism for Kademlia can reduce the lookup latency under churn and, even if the churn rate varies remarkably, the network is flexible enough to provide fine performance.
- FCA-ItswTCM :A Fair Adaptive Labeling Algorithm on Congestion Control Mechanism in the DiffServ–
Compared with other labeling algorithms of sliding window, simulation experiment results show that FCA-ItswTCM has better effects on ensuring fairness between TCP and UDP flows and raising resource utilization and system stabilization.
- Research on Establishing the Early-Warning Index System of Energy Security in China–
The energy security in China is defined from three aspects and the scientific index system of energy security were established in order to using for reference to macro-management of energySecurity in China.
- Intelligent Direction of Arrivals Estimation Based on Unite-direction-character–
An effective dimension-degraded model-building approach based on the DOA matrix decomposing was proposed and had a high estimation precision, so this method has a bright application foreground.
- –
- Fast Fingerprint Extraction Method for Specific Audio–
This paper presents a fast fingerprint extraction method for specific audio that ensures the accuracy of detection and extracts fingerprint rapidly at the same time.
- Research on Subscription Reconfiguration Algorithm in Content-based Publish/Subscribe System–
The algorithm improves subscription forwarding method and forwards the message only on reconfiguration path, which avoids subscription conflict and improves efficiency, and experimental results show the algorithm decreases subscription reconfigured time and its performance suffers little influence from network scale.
- –
- –
- A new central map based on trapdoor of step-wise triangular systems–
The analysis results of the algorithm show that the central map can control the length of key and recover the plaintext from the intermediate ciphertext quickly, which enables its application in constructing a high power of multivariate public key cryptosystems.
- –
- LDPC codes for LTE-Advanced and research on encoding and decoding algorithm–
This paper provided an effective coding and decoding algorithm scheme for hardware design in LTE-Advanced, having some practical value.
- –
- IDMEF-Based Architecture of Large-Scale Cooperative IDS System–
The design and implementation of the cooperative IDS are discussed by the means of Prelude framework and development suite, and the deployment and application of this architecture on CERNET are analyzed.
- –
- Routing Algorithm for Content-based Publish/Subscribe Based on Cyclic Topology–
Analysis shows that the algorithm can completely meet the requirement of P/S system of cyclic topology and guarantees that all the subscribers have subscription matched the event will receive the event without routing loops.
- –
- –
- A Semantics-Supported Publish/Subscribe System–
The experimental results show that the design method of semantics-supported publish/subscribe supports different semantics to a certain extent, and improves the accuracy of matching.
- A New Fast Large-number Modular Multiplication Algorithm Based on Addition–
A new fast calculating AB mod N algorithm based on addition is presented, which introduces pre-calculating by modular N and the computing complexity is reduced largely.
- A New Method Of Processing Blog Posts For Sentiment Classification.–
A feed pipe for conveying sticky bulk material from a bunker outlet vertically downward to a conveyor apparatus wherein the feed pipe is comprised of a plurality of truncated conical sections tapering downwards, the axial length and the diameter of the tubular sections being dimensioned such that arching of the bulk material is avoided.
- On the construction of digital library in higher schools–
The article states the basic measures to construct digital library in higher schools and the meaning and characteristics of digital library.
- –
- Research on Dependency of Flushing of Pending Commit Transaction Log in Kernelized Multilevel Secure Database System–
A trusted log coordinator is introduced to maintain the global dependency of pending commit transactions’ commit log record coordinate the log writing of different DBMS instances to guarantee that the depended commit log records are written out before the depending ones.
- –
- –
- An Efficient Wireless Resource Allocation Based on a Data Compressor Predictor–
A mobility prediction algorithm is proposed for resource allocation and reservation for wireless network and it outperforms Fixed-percent and ExpectedMax in the QoS support effectiveness.
- On Recognition of Handwritten Chinese CharactersBased on Improved BP Networks–
Experimental results demonstrate that this Chinese character recognition system can successfully overcome obstacles out of the complexity of the Chinese characters' structures, and of difficulties to foresee written forms' changes.
- Improved Filtering Method for Medical Images–
An improved filtering method for medical images which can be removed effectively as well as the edges and details of images can be kept well and the shortcoming that burrs appears in images which is filtered using this method is overcomed.
- –
- Build Virtual Classroom of School Network–
This paper introduces methods of build virtualclassroom of school network that makes use of Multicast and VOD, Provide characteristic of CSCL, intelligent networking services.
- –
- –
- The architecture of Multicast in Mobile IP–
An improved architecture to support IP multicast for mobile hosts using mobile IP is provided, comparing it with the introduced mechanism.
- A category of mathematical model for teaching quality assessment–
A category of quantitative analysis method is given for teaching quality assessment by using the non post behavior with the Markov process, which reveals emphatically the progress and variation of whole teaching process.
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.