Is this you? Claim this profile to correct it, add a bio and choose the work people see first.
Claim this profileWorks209 from public data
- LIMO: Less is More for Reasoning546
It is demonstrated that sophisticated mathematical reasoning can emerge with only a few examples and suggested that the threshold for eliciting complex reasoning is not dictated by task complexity but rather by two key factors: the completeness of the model's pre-trained knowledge base and the effectiveness of post-training examples in serving as cognitive templates that guide reasoning.
- Open-Domain Targeted Sentiment Analysis via Span-Based Extraction and Classification230
This work proposes a span-based extract-then-classify framework, where multiple opinion targets are directly extracted from the sentence under the supervision of target span boundaries, and corresponding polarities are then classified using their span representations.
- Reinforced Mnemonic Reader for Machine Reading Comprehension216
The Reinforced Mnemonic Reader for machine reading comprehension tasks, which enhances previous attentive readers in two aspects: a reattention mechanism is proposed to refine current attentions by directly accessing to past attentions that are temporally memorized in a multi-round alignment architecture.
- Read + Verify: Machine Reading Comprehension with Unanswerable Questions164
This work proposes a novel read-then-verify system, which not only utilizes a neural reader to extract candidate answers and produce no-answer probabilities, but also leverages an answer verifier to decide whether the predicted answer is entailed by the input snippets.
- A social network approach to trust management in VANETs131
To the best of the knowledge, this work is the first to introduce information cascading and oversampling to VANETs, and shows that simple voting for decision making leads to oversamplings and gives incorrect results in VANets.
- A Multi-Type Multi-Span Network for Reading Comprehension that Requires Discrete Reasoning99
The Multi-Type Multi-Span Network (MTMSN) is introduced, a neural reading comprehension model that combines a multi-type answer predictor designed to support various answer types with amulti-span extraction method for dynamically producing one or multiple text spans.
- From Yes-Men to Truth-Tellers: Addressing Sycophancy in Large Language Models with Pinpoint Tuning78
This work proposes a novel supervised pinpoint tuning (SPT), where the region-of-interest modules are tuned for a given objective, and demonstrates that SPT significantly mitigates the sycophancy issue of LLMs (even better than SFT).
- Spatio-Temporal Inception Graph Convolutional Networks for Skeleton-Based Action Recognition78
This work designs a simple and highly modularized graph convolutional network architecture for skeleton-based action recognition by repeating a building block that aggregates multi-granularity information from both the spatial and temporal paths.
- Attention-Guided Answer Distillation for Machine Reading Comprehension77
This paper demonstrates that vanilla knowledge distillation applied to answer span prediction is effective for reading comprehension systems and proposes two novel approaches that not only penalize the prediction on confusing answers but also guide the training with alignment information distilled from the ensemble.
- Alterations of Muscle Synergies During Voluntary Arm Reaching Movement in Subacute Stroke Survivors at Different Levels of Impairment72
By comparing the structure of muscle synergies and the similarity of activation coefficients across groups, this paper can validate the increasing activation of pectoralis major muscle and the decreasing activation of elbow extensor of triceps in stroke groups.
- Learning Environmental Sounds with Multi-scale Convolutional Neural Network72
A novel end-to-end network called WaveMsNet is proposed based on the multi-scale convolution operation and two-phase method, which can get better audio representation by improving the frequency resolution and learning filters cross all frequency area.
- Rapid adaptation for deep neural networks through multi-task learning70
The proposed MTL adaptation framework improves the learning ability of the original DNN structure, then enlarge the coverage of the acoustic space to deal with the unseen senone problem, and thus enhance the discrimination power of the adapted DNN models.
- Retrieve, Read, Rerank: Towards End-to-End Multi-Document Reading Comprehension69
RE^3QA is presented, a unified question answering model that combines context retrieving, reading comprehension, and answer reranking to predict the final answer and outperforms the pipelined baseline and achieves state-of-the-art results on two versions of TriviaQA and two variants of SQuAD.
- Recent Trends in Deep Learning Based Open-Domain Textual Question Answering Systems67
This work presents a thorough survey to explicitly give the task scope of open-domain textual QA, overview recent key advancements on deep learning based open- domain textualQA, illustrate the models and acceleration methods in detail, and introduce open- domains textual QC datasets and evaluation metrics.
- DNN-based speech bandwidth expansion and its application to adding high-frequency missing features for automatic speech recognition of narrowband speech65
These techniques not only improve over the traditionally-adopted objective and subjective measures but also reduce the word error rate (WER) from 8.67% when recognizing narrowband speech to 8.26% when recognizing bandwidth-expanded speech.
- Automating psychological hypothesis generation with AI: when large language models meet causal graph63
The results show that combining LLM with machine learning techniques such as causal knowledge graphs can revolutionize automated discovery in psychology, extracting novel insights from the extensive literature.
- 56
- Motor Function Assessment of Upper Limb in Stroke Patients52
The five features proposed in this paper can quantitatively evaluate the motor function of stroke patients which is very useful in the rehabilitation process.
- 52
- 50
- 48
- Electroacupuncture as an adjunctive therapy for motor dysfunction in acute stroke survivors: a systematic review and meta-analyses48
This review provides new evidence for the effectiveness and safety of EA combined with RT and/or CD for PSMD, however, the results should be interpreted cautiously because of methodological weakness and publication bias.
- Exploring Fourier Prior for Single Image Rain Removal47
A two-stage model where the first stage restoring the amplitude of rainy images to clean rain streaks, and the second stage restores the phase information to refine fine-grained background structures to improve the generalization ability of an image rain removal model.
- SNDCNN: Self-Normalizing Deep CNNs with Scaled Exponential Linear Units for Speech Recognition42
The self-normalizing deep CNN (SNDCNN) based acoustic model topology is proposed, by removing the SC/BN and replacing the typical RELU activations with scaled exponential linear unit (SELU) in ResNet-50.
- 41
- 41
- 40
- Revisiting Knowledge Distillation: An Inheritance and Exploration Framework39
This paper proposes a novel inheritance and exploration knowledge distillation framework (IE-KD), in which a student model is split into two parts - inheritance and Exploration, which can be a general technique to improve the student network to achieve SOTA performance.
- Natural history and genetic study of LAMA2-related muscular dystrophy in a large Chinese cohort39
This study provides better understandings of natural history and genotype–phenotype correlations in LAMA2 -related muscular dystrophy, and supports therapeutic targets for future researches.
- Developing and Preliminary Validating an Automatic Cell Classification System for Bone Marrow Smears: a Pilot Study38
An automatic system with integrated scanning hardware and machine learning-based software to perform differential cell count on bone marrow smears to assist diagnosis and provides a useful digital tool in the screening and evaluation of various hematological disorders.
- Improving the Robustness of Electromyogram-Pattern Recognition for Prosthetic Control by a Postprocessing Strategy35
A postprocessing strategy that locks the outputs during the constant contraction to block out occasional misclassifications, upon detecting the motion onset using a threshold, which would facilitate the clinical application of EMG-PR-based prosthetic control.
- A fast method based on multiple clustering for name disambiguation in bibliographic citations35
A novel coarse‐to‐fine framework for name disambiguation which sequentially employs 3 common and easily accessible attributes (i.e., coauthor name, article title, and publication venue) is proposed which outperforms the existing state‐of‐the‐art methods.
- A unified deep framework for peptide–major histocompatibility complex–T cell receptor binding prediction34
Unified deep framework for peptide, MHC and TCR binding prediction to predict the binding between the peptide and the CDR3 of TCR β in general, presented by class I MHCs is proposed.
- Motor impairment evaluation for upper limb in stroke patients on the basis of a microsensor32
Experimental results show that the proposed motor feature indices reflect the real motor impairment status very well, providing objective measurements of a patient’s functional recovery status in the rehabilitation process.
- Text visualization for geological hazard documents via text mining and natural language processing30
The results show that text visualization analysis can be used to identify the types and locations of geological disasters in reports, highlight key information from survey reports as an auxiliary public resource, and more rapidly analyze the key contents of a large number of geological disaster survey reports.
- Argumentation Mining on Essays at Multi Scales30
Experimental results indicate that the proposed multi-scale argumentation mining model outperforms existing models on mining all types of argumentation components and an effective coarse-to-fine argumentation fusion mechanism is proposed to further improve the performance.
- Interactively-Propagative Attention Learning for Implicit Discourse Relation Recognition24
A propagative attention learning model using a cross-coupled two-channel network that yields substantial improvements over the baselines (BiLSTM and BERT).
- Minimizing data redundancy for high reliable cloud storage systems23
This paper proposes a practical and efficient storage allocation scheme, which is proved to be able to minimize the data redundancy and analytically demonstrates that the suggested solution brings several advantages, in particular the reduction of the search space and the acceleration to the computation.
- OlympicArena Medal Ranks: Who Is the Most Intelligent AI So Far?22
This report proposes using an Olympic medal Table approach to rank AI models based on their comprehensive performance across various disciplines, and reveals that there is still a long way to go before achieving superintelligence.
- Meta-Learning Based Knowledge Extrapolation for Temporal Knowledge Graph22
Experimental results on two different TKG extrapolation datasets show that MTKGE consistently outperforms both the existing state-of-the-art models for knowledge graph extrapolation and specifically adapted KGE and TKGE baselines.
- 22
- 21
- 20
- Generalized Semantic Contrastive Learning via Embedding Side Information for Few-Shot Object Detection19
This work proposes a generalized feature learning method for FSOD by leveraging side information, and introduces a side-information guided region-aware masking module that improves sample diversity by removing biased information through counterfactual explanations.
- Incorporating Structured Sentences with Time-enhanced BERT for Fully-inductive Temporal Relation Prediction18
This work extends the fully-inductive setting, where entities in the training and test sets are totally disjoint, into TKGs and takes a further step towards a more flexible and time-sensitive temporal relation prediction approach SST-BERT, incorporating Structured Sentences with Time-enhanced BERT.
- Contextualization of ASR with LLM using phonetic retrieval-based augmentation17
This work proposes a retrievalbased solution to contextualize the LLM: first let the LLM detect named entities in speech without any context, then use this named entity as a query to retrieve phonetically similar named entities from a personal database and feed them to the LLM, and finally run context-aware LLM decoding.
- Recent Advances in Medical Image Processing17
This paper discusses in detail the architecture of a convolution neural network through visualization to help understand its internal working mechanism and several open questions, current trends, and critical challenges faced by medical image processing and artificial intelligence technology.
- Characteristics of Primary Stabbing Headache in a Tertiary Neurological Clinic in China17
PSH was shown to have an onset at middle age with moderate-to-severe intensity attacks localized predominantly within the first division of the trigeminal nerve.
- DaMSTF: Domain Adversarial Learning Enhanced Meta Self-Training for Domain Adaptation16
A new self-training framework for domain adaptation, namely Domain adversarial learning enhanced Self-Training Framework (DaMSTF), which involves meta-learning to estimate the importance of each pseudo instance, so as to simultaneously reduce the label noise and preserve hard examples.
- 16
- Reducing Repair Traffic in P2P Backup Systems16
This article proposes a new class of erasure codes called ER-Hierarchical Codes that combine the advantages of both, Regenerating Codes and Hierarchical codes, that are proposed that significantly reduce the repair traffic.
- Beyond cross-entropy: towards better frame-level objective functions for deep neural network training in automatic speech recognition15
Two approaches for improving the objective function for the deep neural network (DNN) frame-level training in large vocabulary continuous speech recognition (LVCSR) are proposed, which considers jointly minimizing the cross-entropy and maximizing the log posterior ratio between the target senone and the most competing one.
- SASGBC: Improving Sequence Labeling Performance for Joint Learning of Slot Filling and Intent Detection14
A new joint learning model called SASGBC is proposed that combines sequence labeling with deep bidirectional Transformer and achieves state-of-the-art performance compared to several competitive baseline models.
- 13
- Effective segmentations in white blood cell images using $$\epsilon $$ ϵ -SVR-based detection method13
Two applications are carried out, including an adaptive thresholding algorithm based on WBC detection for nucleus segmentation from images and target detection to lessen the users’ interaction for automatic cytoplasm segmentation.
- Towards an Extensible and Secure Cloud Architecture Model for Sensor Information System12
This architecture fully utilizes the virtualization technology and cloud service functions and builds an extensible cloud platform environment supplying scalable, flexible, reliable, and secure sensor information services, which aims to solve the challenges encountered by the present sensor information system.
- On Reputation and Data-centric Misbehavior Detection Mechanisms for VANET12
This thesis points out several limitations of reputation trust management schemes for VANET, and proposes a solution to overcome the problem of information cascading and oversampling, which commonly arise in social networks.
- A Canonicalization-Enhanced Known Fact-Aware Framework For Open Knowledge Graph Link Prediction11
This work proposes a Canonicalization-Enhanced Known Fact-Aware (CEKFA) framework that boosts link prediction performance through sparsity reduction of RPs and triples, and proposes a similarity-driven RP canonicalization method to reduce RPs' sparsity by sharing knowledge of semantically similar ones.
- Scenario-independent Uncertainty Estimation for LLM-based Question Answering via Factor Analysis10
A plug-and-play scenario-independent framework to enhance unsupervised UE in LLMs by removing scenario-related noise and focusing on semantic information is introduced and leverages only the existing UE methods' outputs.
- Personalization of CTC-Based End-to-End Speech Recognition Using Pronunciation-Driven Subword Tokenization10
This work presents a novel method for generating additional subword tokenizations for personal entities from their pronunciations and shows that using this technique in combination with two established techniques, contextual biasing and wordpiece prior normalization, it is able to achieve personal named entity accuracy on par with a competitive hybrid system.
- 10
- Using a Penalty-based Loss Re-estimation Method to Improve Implicit Discourse Relation Classification10
A penalty-based loss re-estimation method is proposed to regulate the attention learning process, integrating penalty coefficients into the computation of loss by means of overstability of attention weight distributions.
- A Bounded Scheduling Method for Adaptive Gradient Methods10
Experimental results show that the proposed bounded scheduling algorithm for Adam can eliminate the generalization gap between Adam and SGD, meanwhile maintaining a relative high convergence rate during training.
- 9
- Hierarchical Shared Encoder With Task-Specific Transformer Layer Selection for Emotion-Cause Pair Extraction9
A multi-task learning framework named Hierarchical Shared Encoder with Task-specific Transformer Layer Selection (HSE-TTLS) is proposed, which outperforms state-of-the-art baseline methods by at least 1.56% on the F1 score.
- Pegasus: a distributed and load-balancing fingerprint identification system9
Results show that Pegasus can reduce the time cost by 70% during the feature extraction procedure and balances the difference of access load among front-end mongos nodes to less than 5%.
- 9
- Science Navigation Map9
An online tool for literature navigation, filtering, and interactive data mining, named Science Navigation Map (SNM), which integrates information from online paper repositories, citation databases, etc, and proposes a multi-view non-negative matrix factorization, which can make better use of complicated multi-wise relationships in papers.
- 9
- Acoustic Model Fusion For End-to-End Speech Recognition8
A significant reduction in the word error rate is achieved, with an impressive drop of up to 14.3% across varied test sets, and it is discovered that this AM fusion approach is particularly beneficial in enhancing named entity recognition.
- 7
- Research on Global Navigation Operations for Rotary Burying of Stubbles Based on Machine Vision7
The algorithm adopts the Hough matrix to extract the characteristic points of the straw boundaries and remove the redundancies, and then reduces the influence of noise points caused by different straw shapes using the RANSAC algorithm to improve the accuracy of the starting route extraction.
- 7
- 7
- Accelerating SGD using flexible variance reduction on large-scale datasets7
A new flexible variance reduction method for SGD, named FVR-SGD, which can reduce memory overhead and speed up the convergence using flexible subset without extra computation is proposed, and linear convergence rate can be guaranteed while using flexible variance reduce.
- Improving Deep Neural Network Based Speech Synthesis through Contextual Feature Parametrization and Multi-Task Learning7
The bidirectional recurrent neural network architecture with long short term memory (BLSTM) units is adopted and trained with multi-task learning for DNN-based speech synthesis and the quality of synthesized speech has been improved.
- 6
- Meta-Tsallis-Entropy Minimization: A New Self-Training Approach for Domain Adaptation on Text Classification6
Theoretically, the convergence of the meta-learning algorithm in MTEM is proved and the effectiveness of MTEM in achieving domain adaptation is analyzed.
- Teaching What You Should Teach: A Data-Based Distillation Method6
The "Teaching what you Should Teach" strategy is introduced into a knowledge distillation framework, and a data-based distillation method named "TST" is proposed that searches for desirable augmented samples to assist in distilling more efficiently and rationally.
- 6
- 6
- 6
- Multitask Learning for Chinese Named Entity Recognition6
A character-level tagger for Chinese NER, which incorporates multitask learning, self-attention and multi-step training methods to exploit richer features and further improve the model performance is described and evaluated.
- Deadline-Oriented Task Scheduling for MapReduce Environments6
The results show that Dart can not only effectively meet the deadline but also process near-maximal data volumes even when the deadline is set to be extremely small and limited resources are allocated.
- Towards a delivery scheme for speedup of data backup in distributed storage systems using erasure codes6
This work proposed a fast backup scheme in distributed systems based on general erasure coding that can achieve minimum backup time by solution for classical quadratic programming problem and shows how this process can be modeled and derive a formula about the final backup time.
- 6
- White Cloud or Black Cloud: Opportunity and Challenge of Spectrum Sharing on Cloud Computing6
This paper discusses the convergence of spectrum sharing and cloud computing from several aspects including model, advantages and challenges, and a spectrum sharing model based on cloud computing (SSC) will be introduced.
- SolidTrack: A Novel Method for Robust and Reliable Multi-Pedestrian Tracking5
This study analyzes the sensitivity of state-of-the-art TBD-based trackers to variations in the high-confidence threshold and introduces SolidTrack, an enhanced tracking algorithm integrating SOLIDER-ReID, a transformer-based re-identification model that exhibits superior robustness across varying threshold settings.
- 5
- 5
- Exploring Retraining-free Speech Recognition for Intra-sentential Code-switching5
This work presents a prototype research code-switching speech recognition system that leverages existing monolingual acoustic and language models, i.e., no ad hoc training is needed and handles accented foreign pronunciations better than techniques based on human labeling.
- 5
- 4
- SimCroP: Radiograph Representation Learning with Similarity-Driven Cross-Granularity Pre-training4
A Similarity-Driven Cross-Granularity Pre-training (SimCroP) framework on chest CTs is proposed, which combines similarity-driven alignment and cross-granularity fusion to improve radiograph interpretation and outperforms both cutting-edge medical self-supervised learning methods and medical vision-language pre-training methods.
- A cause fusion framework with information bottleneck for conversational causal emotion entailment4
This work proposes a novel framework called Cause Fusion framework with Information Bottleneck (CFIB), which takes into account the inter-cause relationships and places greater emphasis on cause-related features and reduces the impact of irrelevant emotional information.
- Plot-aware transformer for recommender systems4
This paper proposes the "Plot-Aware Transformer" model (PAT), currently the best natural language processing module, to model the process of "user-movie" rating interaction and test the model on several movie datasets and demonstrated that the model is competitive.
- Assessment of the operation status of electronic prescription at community pharmacies in Chengdu, China: a simulated patient study4
The electronic prescription services mode for community pharmacies in Chengdu provides a convenient drug purchase process but remains some problems, for example, prescribing drugs without a prescription and services provided by pharmacists was poor, etc.
- 4
- 4
- 4
- Scoliosis Orthopedic Surgery Combined With Nusinersen Intrathecal Injection Significantly Improved the Outcome of Spinal Muscular Atrophy Patient: A Case Report4
Scoliosis orthopedic surgery combined with Nusinersen intrathecal injection is an effective treatment for SMA patients with scoliosis.
- Spatial Attention Network for Head Detection4
A novel spatial attention network (SAN) which adopts the saliency module to exploit the environmental information beyond the proposal which is ignored in the Faster-RCNN is proposed.
- 4
- Rethinking the Hidden Risk of Reranking: Achieving Risk-aware Reranking with Information Gain for RAG with LLMs3
A risk-aware reranking method for RAG with LLMs is proposed, which balances the risk and benefit during reranking, and reveals a counterintuitive finding: though rerankers improve the proportion of GD, they also significantly increase the proportion of harmful documents in top-ranked retrieved documents.
- 3
- 3
- FUR-DETR: A Lightweight Detection Model for Fixed-Wing UAV Recovery3
A lightweight visual detection model based on transformer architecture based on transformer architecture is proposed, which reduces the parameter size and computational load by 43.2% and 58% while maintaining detection accuracy comparable to the original RT-DETR in three datasets, and an adaptive multiscale feature pyramid network (AMFPN) is designed.
- 3
- Exploration, fusion, and refinement: a multivariate features interaction network for visual camouflaged detection3
The proposed MFINet is evaluated on four challenging COD benchmark datasets using four widely used metrics and demonstrates that the proposed method is superior to other existing cutting-edge methods for camouflaged object detection.
- Study on the Detection Mechanism of Multi-Class Foreign Fiber under Semi-Supervised Learning3
The proposed method introduces CBAM to address the issue of the missed detection and false detection of small-sized cotton foreign fibers against complex backgrounds and designs a multiscale feature information extraction network, SPPFCSPC, which improves its ability to generalize to fibers of different shapes.
- 3
- 3
- 3
- 3
- [Preliminary study of robot-assisted ankle rehabilitation for children with cerebral palsy].3
The robot assisted ankle-foot rehabilitation system can improve the contracture and soft tissue compliance of cerebral palsy children's ankle plantar flexors and may be suitable for application in the rehabilitation of children with cerebral palsY.
- 3
- An article level metric in the context of research community3
Experimental results show the novel metric is effective and robust in assessing papers, and it is consistent, practical and reasonable to calculate the impact of the journals and scholars under relatively fair criteria.
- 3
- Robust Redundancy Scheme for the Repair Process: Hierarchical Codes in the Bandwidth-Limited Systems3
A near-optimal erasure code named Hierarchical Codes, has been proposed that can significantly reduce the repair traffic by reducing the number of nodes participating in the repair process, which is referred to as the repair degreed.
- 3
- LTOS: Layout-controllable Text-object Synthesis via Adaptive Cross-attention Fusions2
This paper introduces a unified task called layout-controllable text-object synthesis (LTOS), which merges text rendering and layout-based object generation, and proposes a self-adaptive cross-attention fusion that forces the image generation to attend more to important text spans.
- Spectral Information Divergence-Driven Diffusion Networks for Hyperspectral Target Detection2
A Spectral Information Divergence-driven Diffusion Network model (SID-DN) for hyperspectral target detection is proposed, which significantly enhances detection robustness in complex scenes by decoupling background distribution modeling from target detection.
- 2
- 2
- Temporal Extrapolation and Knowledge Transfer for Lifelong Temporal Knowledge Graph Reasoning2
This work forms lifelong TKG reasoning as a temporal-path-based reinforcement learning (RL) framework, which adds temporal displacement into the action space of RL to extrapolate for the future and further proposes a temporal-rule-based reward shaping to guide the training.
- Structure Enhanced Path Reasoning for Knowledge Graph Completion2
Experimental results show that this approach apparently outperforms state-of-the-art methods on fact prediction task and relation prediction task, and extensive experiments illustrate the benefits of enhancing path reasoning by exploiting structure information from entity relational contexts and the dual path encoding method.
- 2
- A BERT-Based Semantic Matching Ranker for Open-Domain Question Answering2
This paper proposes a simple but effective BERT-based semantic matching ranker to compute the semantic similarity between the paragraph and given question, in which three different representation aggregation functions are explored.
- MFM: A Multi-level Fused Sequence Matching Model for Candidates Filtering in Multi-paragraphs Question-Answering2
Experiments show that the proposed multi-level fused sequence matching model can outperform recent models not only on the filtering candidates in multi-paragraphs QA task but also on the sequence matching task.
- 2
- 2
- 2
- A Packet Scheduling Method for Multimedia QoS Provisioning2
Early Flow Discard EFD is proposed, a new size-based scheduler, which keeps the salient properties of state-of-the-art size based schedulers like LARS, with a bounded memory requirement, and is demonstrated to be efficient and an interesting solution to the so-called bufferbloat phenomenon.
- 2
- Repairing Multiple Data Losses by Parallel Max-min Trees Based on Regenerating Codes in Distributed Storage Systems2
Simulation experiment shows that multiple parallel max-min trees reduce total regeneration time for multiple data losses significantly, and thus enhance system reliability, compared with existing regeneration scheme.
- Partial Clones for Stragglers in MapReduce2
A new approach for mitigating stragglers in MapReduce is put forward, name Hummer, which starts task clones only for high-risk delaying tasks and can decrease the job delaying risk with fewer resources consumption.
- Class-Patch Similarity Weighted Embedding for Few-Shot Infrared Image Classification1
The Class Patch Similarity Weighted Embedding (CPSWE) framework for few-shot infrared target classification outperforms existing few-shot learning methods, achieving significant improvements in classification accuracy on infrared image datasets with limited labeled samples.
- 1
- 1
- 1
- 1
- 1
- 1
- 1
- 1
- Research on Att-NFM Recommendation Algorithm Based on Attention Mechanism1
This work proposes the Att-NFM model to learn attention coefficients in neural networks to characterize the importance between features and shows that the evaluation metrics are improved by 3% and 2%, respectively.
- WCSR: A Robotic Process Automation Data Collection Method Based on Web Content Structure Recognition1
- 1
- 1
- 1
- 1
- 1
- FVR-SGD: A New Flexible Variance-Reduction Method for SGD on Large-Scale Datasets1
A new flexible variance reduction method for SGD is proposed, named FVR-SGD, which can reduce memory overhead and speedup the convergence using flexible subset size without extra operation, and demonstrated that it outperforms contemporary SVRG algorithm.
- 1
- 1
- Reducing Insertion Time in LRC-based Cloud Storage Systems1
This work proposes a new method combined global and local optimization to schedule the in-network redundancy generation process and shows that this method reduce insertion time up to 18% compared with the best heuristic scheduling algorithm.
- 1
- 1
- Improving Transmission Availability for Wireless Cognitive Networks: SkyChannel Query Processing1
This work proposes an improving transmission availability algorithm— SkyChannel Query, which can find the non dominated channels and only transmit information of these channels and the experimental results show the advantages of the algorithm in terms of saving computing resource and improve transmission availability.
- An Application-Oriented Network Model for Wireless Sensor Networks1
An application-oriented general network model (AGNM) for RERM is proposed and the experimental results show that AGNM can approximately characterize the real situation of WSNs, therefore, the AGNM provides a good foundation for RerM research.
- 1
- –
- –
- –
- 3-Hydroxysterol Δ24-Reductase Promotes Ovarian Cancer Progression by Activating the TGF-1/Smad2/3 Signaling Pathway–
DHCR24 contributes to ovarian cancer progression by upregulating the TGF-β1 pathway, highlighting its potential as a therapeutic target in ovarian cancer.
- Deciphering Virulence Factors of Hyper-Virulent Pseudomonas aeruginosa Associated with Meningitis.–
Light is shed on the virulence factors of P. aeruginosa involved in meningitis by identifying its wide-spectrum resistance and downregulation of the blood-brain barrier-associated proteins Claudin-5 and Occludin by A584.
- –
- Freeway vehicle detection on improved YOLOv7–
Improved data augmentation method is applied to simulated the special environment interference, and SimAM attention mechanism module is inserted into YOLOv7, and the optimal adding position is obtained by discussing the performance of different attention composition.
- Research on reliability analysis of underwater constant current supply system in mesh topology model–
- –
- –
- –
- Adaptive resource scaling of service mesh in cloud computing–
This paper proposes an adaptive resource scaling approach for service mesh that leverages practical method based on a feedback loop that continuously monitors the proxies’ CPU utilization and CPU throttling to predict its future resource requirements.
- The application of convolutional neural networks in face detection–
The comparison between various deep learning-based methods in terms of the performance indicated that there is still no high robustness solution to all problems.
- –
- Deep-to-Bottom Weights Decay: A Systemic Knowledge Review Learning Technique for Transformer Layers in Knowledge Distillation–
A deep-to-bottom weights review mechanism to knowledge distillation, which fuses teacher-side information 015 taking each layer’s difficulty level into consid-016 eration and results show that the review approach is able to outperform other existing techniques.
- Fixed-Size Objects Encoding for Visual Relationship Detection–
The proposed fixed-size object encoding method, FOE-VRD, works well on both predicate classification and relationship detection tasks, especially on the situation of zero-shot detection.
- –
- –
- –
- –
- –
- Open-Domain Textual Question Answering Systems–
This chapter provides an overview of open-domain QA systems, then introduces models on paragraph ranking, candidate answer extraction, and answer selection.
- –
- –
- –
- –
- –
- Exploring how Computer-Assisted Language Learning (CALL) supports English acquisition from the perspectives of international students with Asian–
It is suggested that for the Asian students who are in the process of transition from a traditional teacher centred approach to a student Centred approach, more assistance and guidelines are needed for them to gradually develop their computer skills, independent learning skills and self-motivational skills to benefit from using CALL.
- Parallel Data Regeneration Based on Multiple Trees with Network Coding in Distributed Storage System–
This paper considers to use multiple edge-disjoint trees to parallel regenerate the lost data, and proposes an approximate construction algorithm with polynomial time complexity for the optimal multiple regeneration trees.
- Solution method of simplified classification power network model for Preserving Key Branches–
An algorithm based on power transfer distribution factor for computing the reduced networks flows is presented and an improved centroid sorting technology is applied for obtaining the optimal reduced model.
- –
- Tree-Structured Parallel Regeneration Based on Regenerating Codes for Multiple Data Losses in Distributed Storage Systems–
A tree-structured parallel regeneration scheme based on regenerating codes (TPRORC) for multiple data losses in distributed storage systems is proposed, which reduces the regeneration time significantly, compared to other regular regeneration schemes.
- –
- Application Study about OTSU Optimal Threshold Method in Cotton Seedling Recognition–
It is explained that the method of using visual picture processing technology to recognize cotton seedling is feasible.
- Optimizing Hidden Markov Model with Baum-Welch Algorithm for Vehicle Driver's Intention Recognition–
The Baum-Welch (B-W) algorithm is introduced to optimize and update the parameters of the established HMM to solve the problem of vehicle driver’s intention recognition.
- Semantic-Driven Information Recommendation System–
SIRSCA mainly includes four modules: semantics representation of foundation data and user preference informations; indexing mechanism of massive semantic informations under cloud architecture; recommendation approaches based on semantic computation theory; and technologies of dynamic migration under cloud Architecture.
- –
- –
- Queueing Analysis of the Decoding Process for Intra-session Network Coding with Random Linear Codes–
A multi-channel batch service queueing system with control feedbacks is built to describe the decoding process of intra-session network coding with random linear codes and try to answer several fundamental questions, including for example, how to analyze braking redundancy.
- Hierarchical Codes in Bandwidth-Limited Distributed Storage Systems–
A near-optimal erasure code named Hierarchical Codes is proposed that can significantly reduce the repair traffic by reducing the number of nodes participating in repair, which is referred to as repair degree d.
- Rack-Aware Architecture for Reliable Cloud Systems–
A rack-aware architecture for cloud storage systems based on fault-tolerant mechanism for large-scale network storage systems is proposed.
- Queueing Analysis of Spectrum Trading for Wireless Cognitive Radio Networks–
A spectrum trading model based on queueing theory is considered for getting the most profit from the spectrum trading and the relationship between the spectrum server and the spectrum demander is analyzed and the problem of server uncertainty is discussed.
- Effect of Rack-Aware Placement in Cloud Storage System–
This model aims to tell the effect of rack-aware placement for cloud storage systems by simplifying and analyzing the data reliability by a markov model, which includes the discussion of correlated failures and rack- Aware placement.
- –
- –
- –
- –
- Study and implementation of broadband wireless IP network adapter–
With CCK and advanced broadband RF technologies, a new design scheme for the adapter is presented and implemented to transmit high speed and burst data reliably in the 2 4?GHz band.
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 2026-10-15. 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.