Is this you? Claim this profile to correct it, add a bio and choose the work people see first.
Claim this profileWorks286 from public data
- Gated Residual Recurrent Graph Neural Networks for Traffic Prediction261
A novel end-to-end multiple Res- RGNNs framework, referred to as “MRes-RGNN”, for traffic prediction is proposed, which outperforms state-of-the-art methods significantly.
- Multiple convolutional neural networks for multivariate time series prediction238
Tests on two large real-world datasets, show that the proposed model, called Multiple CNNs, has a strong advantage over other time series prediction methods.
- Citywide Traffic Flow Prediction Based on Multiple Gated Spatio-temporal Convolutional Neural Networks235
An end-to-end framework, multiple gated spatio-temporal CNNs (MGSTC), for citywide traffic flow prediction and demonstrates that MGSTC outperforms other state-of-the-art baselines.
- 181
- Tumor‐derived exosomal lncRNA GAS5 as a biomarker for early‐stage non‐small‐cell lung cancer diagnosis176
Exo‐GAS5 may function as an ideal noninvasive serum‐based marker for identifying patients with early NSCLC and could be used to distinguish patients with Stage INSCLC with an AUC of 0.822.
- 127
- A hierarchical deep convolutional neural network and gated recurrent unit framework for structural damage detection122
This work proposes a novel Hierarchical CNN and Gated recurrent unit (GRU) framework to model both spatial and temporal relations, termed as HCG, for structural damage detection, where CNN is utilized to model the spatial relations and the short-term temporal dependencies among sensors, while the output features of CNN are fed into the GRU to learn the long- term temporal dependencies jointly.
- 117
- 115
- Hydrothermal fabrication of porous hollow hydroxyapatite microspheres for a drug delivery system111
The PHHMs displayed a high drug loading capacity and pH-responsive sustained-controlled drug release behavior when the authors used doxorubicin hydrochloride (DOX) as a loading drug and the controlled drug release system showed a high ability to kill cancer cells and less damage to normal cells.
- 110
- 90
- WiFi-Based Indoor Robot Positioning Using Deep Fuzzy Forests89
A lightweight indoor robot positioning system which operates on cost-effective WiFi-based received signal strength (RSS) and could be readily pluggable into any existing WiFi network infrastructures is designed and a novel deep fuzzy forest is proposed to inherit the merits of decision trees and deep neural networks within an end-to-end trainable architecture.
- TitAnt78
The TitAnt is introduced, a transaction fraud detection system deployed in Ant Financial, one of the largest Fintech companies in the world, able to predict online real-time transaction fraud in mere milliseconds.
- 76
- DS-AL: A Dual-Stream Analytic Learning for Exemplar-Free Class-Incremental Learning72
Empirical results demonstrate that the DS-AL, despite being an exemplar-free CIL technique, delivers performance comparable with or better than that of replay-based methods across various datasets, including CIFAR-100, ImageNet-100 and ImageNet-Full.
- 71
- Multiple local 3D CNNs for region-based prediction in smart cities69
A novel region-based information extraction mechanism and an end-to-end multiple spatial-temporal dependency learning structure are designed for local regions and the superior performance of the proposed method over the exisiting state-of-the-art methods is demonstrated.
- 69
- 69
- 64
- Three-dimensional graphene biointerface with extremely high sensitivity to single cancer cell monitoring63
Simultaneous CCD recording and electrical signal extraction from the entrapped single cell on the graphene biointerface enabled to investigate multidimensional cell-electrode interactions and predict cancerous stage and pathology.
- miR-151a-3p-rich small extracellular vesicles derived from gastric cancer accelerate liver metastasis via initiating a hepatic stemness-enhancing niche62
A previously unidentified regulatory axis initiated by sEV-miR-151a-3p that establishes a hepatic stemness-permissive niche to support GC-LM is revealed.
- Eraser: Jailbreaking Defense in Large Language Models via Unlearning Harmful Knowledge61
A novel defense method called Eraser is proposed, which mainly includes three goals: unlearning harmful knowledge, retaining general knowledge, and maintaining safety alignment, which can significantly reduce the jailbreaking success rate for various attacks without compromising the general capabilities of the model.
- Modeling Temporal Patterns with Dilated Convolutions for Time-Series Forecasting60
The proposed HyDCNN is a novel hybrid framework based on fully Dilated CNN for time-series forecasting tasks, in which the proposed position-aware dilated CNNs are utilized to capture the sequential non-linear dynamics and an autoregressive model is leveraged to captured the sequential linear dependencies.
- 58
- A novel energy cooperation framework for community energy storage systems and prosumers56
An energy cooperation platform is introduced as a manager to interact with all players and the distribution system operator (DSO), improving the interaction efficiency and having better scalability and a profit-sharing mechanism based on the asymmetric Nash bargaining model is designed.
- Covalent Modifications of the Bacteriophage Genome Confer a Degree of Resistance to Bacterial CRISPR Systems54
Evidence is presented that genome modifications might also have been selected to counter, broadly, the CRISPR-Cas systems, an adaptive bacterial defense mechanism, and that the defense-counterdefense mechanisms exhibited by bacteria and phages, while affording protection against one another, also provide evolutionary benefits for both.
- Accelerating Attention Mechanism on FPGAs based on Efficient Reconfigurable Systolic Array48
A novel method called Four inputs Processing Element (FPE) to double the computation rate of the data-aware systolic array (SA) and make it efficient and load balance and the design is evaluated on a Xilinx Alveo U250 card, and the proposed architecture achieves 51.3× improvement in latency, and 54.9× energy savings compared to CPU and GPU.
- 47
- 46
- Molecular Dynamics and Machine Learning in Catalysts46
This review provides a comprehensive summary of recent developments in molecular dynamics, including ab initio molecular dynamics and reaction force-field molecular dynamics.
- Engineering T4 Bacteriophage for In Vivo Display by Type V CRISPR-Cas Genome Editing43
It is found that the type V CRISPR-Cas12a system, which shows efficient cleavage of ghmC-modified genome when compared to the type II CRISpr-Cas9 system, can be synergistically employed to generate recombinant T4 phages.
- 43
- Bacteriophage T4 Vaccine Platform for Next-Generation Influenza Vaccine Development42
Proof-of-concept for the development of next-generation influenza vaccine using T4 phage virus-like particle (VLP) platform is provided and it is shown that more than 1,280 M2e molecules can be assembled on a 120×90 nanometer phage capsid to form T4-M2e VLPs, which are highly immunogenic and induced complete protection against influenza virus challenge without any addition adjuvant.
- Extreme Learning Machine and Its Applications in Big Data Processing42
This chapter reviews ELM theories and some important variants, and describes parallel ELM algorithms based on MapReduce and Spark in detail, and shows some practical applications of the ELM for big data.
- Campus-Scale Mobile Crowd-Tasking41
This paper designs, develops, develops and experiment with a real-world mobile crowd-tasking platform, called TA$Ker, and provides an in-depth analysis of labor supply, worker behavior & task selection preferences and the efficacy of push-based approaches that recommend tasks based on predicted movement patterns of individual workers.
- IART: Intent-aware Response Ranking with Transformers in Information-seeking Conversation Systems40
This paper analyzes user intent patterns in information-seeking conversations and proposes an intent-aware neural response ranking model “IART”, which refers to “Intent-Aware Ranking with Transformers”.
- 38
- Space-Time Aware Behavioral Topic Modeling for Microblog Posts37
This paper presents a system that combines snapshot isolation (SI) replication with virtualization to provide a flexible solution for running unmodified databases in the cloud while taking advantage of the opportunities cloud architectures provide.
- 37
- CPCM: Contextual Point Cloud Modeling for Weakly-supervised Point Cloud Semantic Segmentation36
A simple yet effective Contextual Point Cloud Modeling (CPCM) method that consists of two parts: a region-wise masking (Region-Mask) strategy and a contextual masked training (CMT) method that disentangles the learning of supervised segmentation and unsupervised masked context prediction for effectively learning the very limited labeled points and mass unlabeled points, respectively.
- A two-stage attention aware method for train bearing shed oil inspection based on convolutional neural networks35
A two-stage attention aware method to recognize defections of bearing shed oil, based on convolutional neural networks, which achieves higher robustness and accuracy in recognizing defect areas with small oil spot and a better segmentation performance is achieved, compared to traditional UNet and related state-of-art approaches.
- Multi-task cascade deep convolutional neural networks for large-scale commodity recognition34
Through extensive experiments and comparative analyses of commodity detection in supermarkets and stores of Jinzhou city, the performance of MTCD-CNN has proved to be superior to other advanced methods, indicating that the proposed method has effectively solved the problem of excessive similarity of confusingly similar categories.
- Multi-rater Prism: Learning self-calibrated medical image segmentation from multiple raters33
A novel neural network framework called Multi-rater Prism (MrPrism) is proposed to learn medical image segmentation from multiple labels and outperforms state-of-the-art (SOTA) methods for a wide range of medical image segmentation tasks.
- Reinforced History Backtracking for Conversational Question Answering32
This paper proposes a reinforcement learning based method to capture and backtrack the related conversation history to boost model performance and considers both immediate and delayed rewards to guide the reinforced backtracking policy.
- Leveraging master-slave OpenFlow controller arrangement to improve control plane resiliency in SD-EONs32
Experimental results indicate that the proposed system can restore services in both the data and control planes of SD-EON jointly while maintaining relatively good scalability.
- TRACCS: Trajectory-Aware Coordinated Urban Crowd-Sourcing32
Credible computationally-efficient heuristics are developed to address the problem of large-scale mobile crowdtasking, where a large pool of citizen crowd-workers are used to perform a variety of location-specific urban logistics tasks.
- 30
- Fate modeling of mercury species and fluxes estimation in an urban river29
The QWASI (Quantitative Water-Air-Sediment Interaction) model was selected to estimate the levels of total mercury and three mercury species in water and sediment, and was used to quantify the fluxes of mercury at water/air and sediment/water interfaces of an urban river.
- 28
- 28
- 28
- 28
- Locally Connected Deep Learning Framework for Industrial-scale Recommender Systems28
A locally connected deep learning framework for recommender systems is proposed, which reduces the complexity of deep neural network (DNN) by two to three orders of magnitude and is extended using the idea of the recently proposed Wide&Deep model.
- 28
- Prediction of Air Quality Based on KNN-LSTM26
It is suggested that the prediction accuracy of the hybrid prediction model constructed in this paper is acceptable in terms of the space-time correlation, and it could be an alternative for further application of air quality prediction.
- 25
- Characterizing Membership Privacy in Stochastic Gradient Langevin Dynamics25
It is suggested that the SGLD algorithm can not only reduce the information leakage but also improve the generalization ability of the DNN models in real-world applications.
- Impacts of sanitation upgrading to the decrease of fecal coliforms entering into the environment in China25
The calculation results indicated that, the current FC release is disparate among regions, and the human releases in the rural regions were dominant, accounting for over 90% of the total human releases.
- CD39+ Regulatory T Cells Attenuate Lipopolysaccharide-Induced Acute Lung Injury via Autophagy and the ERK/FOS Pathway24
The results suggested that the adoptive transfer of CD39+ Tregs may provide a novel method for the clinical prevention and treatment of ARDS.
- 24
- 24
- Hierarchical and Complex System Entropy Clustering Analysis Based Validation for Traditional Chinese Medicine Syndrome Patterns of Chronic Atrophic Gastritis23
Combination of qi deficiency, qi stagnation, blood stasis, phlegm turbidity, heat, yang deficiency, and yin deficiency may play a crucial role in CAG pathogenesis.
- DGSLN: Differentiable graph structure learning neural network for robust graph representations22
This work proposes a novel differentiable graph structure learning neural network (DGSLN), which learns suitable graph structures for GNNs and develops a hybrid loss function to ensure the quality of learned graphs.
- T4 bacteriophage nanoparticles engineered through CRISPR provide a versatile platform for rapid development of flu mucosal vaccines22
It is shown that antigen-decorated nanoparticle generated through CRISPR engineering of T4 bacteriophage can serve as a universal platform for the rapid development of mucosal vaccines and provide a versatile T4 platform that can be customized to rapidly develop mucosal vaccine against future emerging epidemics.
- Functional engineering strategies of 3D printed implants for hard tissue replacement22
This review focuses on the advances of functional engineering strategies for 3D printed biomaterials in hard tissue replacement, introducing 3D printing technologies, properties of printing materials (metals, ceramics and polymers) and typicalfunctional engineering strategies utilized in the application of bone, cartilage and joint regeneration.
- AntProphet: an Intention Mining System behind Alipay's Intelligent Customer Service Bot22
An intention mining system that takes care of more than 90% of the customer service demands in the Alipay APP and resolves most of the users' problems on the spot, thus significantly reduces the burden of manpower.
- Online Analytic Exemplar-Free Continual Learning With Large Models for Imbalanced Autonomous Driving Task20
The AEF-OCL solves the OCL problem by recursively calculating the analytical solution, ensuring an equalization between the continual learning and its joint-learning counterpart, and works without the need to save any used samples (i.e., exemplar-free).
- 20
- Characterization and in vitro biological evaluation of mineral/osteogenic growth peptide nanocomposites synthesized biomimetically on titanium20
Results presented here indicated that the mineral/OGP nanocomposites formed on titanium substrates had the potential for applications in dental implants.
- Research on Improved Multi-Channel Image Stitching Technology Based on Fast Algorithms19
This study proposes a multi-channel image stitching approach based on fast image registration and fusion algorithms, which enhances the stitching effect on the basis of fast algorithms, thereby augmenting its potential for deployment in real-time applications.
- 19
- 18
- 18
- 18
- Semi-Supervised Self-Taught Deep Learning for Finger Bones Segmentation18
This study focuses on segmenting finger bones within a newly introduced semi-supervised self-taught deep learning framework which consists of a student network and a stand-alone teacher module.
- 17
- 16
- 16
- 16
- 16
- 15
- 14
- 14
- 14
- Weakly-Supervised Open-Retrieval Conversational Question Answering14
This work introduces a learned weak supervision approach that can identify a paraphrased span of the known answer in a passage of a passage in the open-retrieval ConvQA setting under a weak supervision setting.
- 14
- Scalable Urban Mobile Crowdsourcing14
This article explicitly allow multiple routine routes to be probabilistically associated with each worker, and formulation of the problem is formulated as an integer linear program whose goal is to maximize the expected total utility achieved by all workers.
- GSDNet: Revisiting Incomplete Multimodality-Diffusion Emotion Recognition from the Perspective of Graph Spectrum13
A novel Graph Spectral Diffusion Network (GSDNet), which utilizes a low-rank score-based diffusion model to map Gaussian noise to the graph spectral distribution space of missing modalities and recover the missing data according to its original distribution.
- Transferability Bound Theory: Exploring Relationship between Adversarial Transferability and Flatness13
TPA is proposed, a Theoretically Provable Attack that optimizes a surrogate of the derived bound to craft adversarial examples that can craft more transferable adversarial examples compared to state-of-the-art baselines.
- A parallel approximate SS-ELM algorithm based on MapReduce for large-scale datasets13
A parallel approximate SS-ELM Algorithm based on MapReduce based on the approximate adjacent similarity matrix (AASM) algorithm, which leverages the Locality-Sensitive Hashing (LSH) scheme to calculate the approximate neighboring similarity matrix, thus greatly reducing the complexity and occupied memory.
- Aspect Extraction from Product Reviews Using Category Hierarchy Information13
An LDA-based generative topic model inducing the two-layer categorical information (CAT-LDA) is proposed, to balance the aspects of both a parent category and its child categories.
- 13
- AIR: Analytic Imbalance Rectifier for Continual Learning12
An analytic imbalance rectifier (AIR) algorithm for real-world CL is proposed that addresses class imbalance with an analytic reweighting module (ARM) that calculates a reweighting factor for each class in the loss function to equalize total sample weights across classes.
- 12
- REAL: Representation enhanced analytic learning for exemplar-free class-incremental learning11
A representation-enhanced analytic learning (REAL) is proposed, which achieves state-of-the-art performance on CIFAR-100, ImageNet-100 and ImageNet-1k benchmarks, outperforming exemplar-free methods and rivaling exemplar-based approaches.
- 11
- Non-contrast MRI of Inner Ear Detected Differences of Endolymphatic Drainage System Between Vestibular Migraine and Unilateral Ménière's Disease11
Based on non-enhanced MRI-visualized measurement, anatomical variables with regard to the endolymphatic drainage system differed significantly between the patients with VM and those with unilateral MD.
- Vascular loops in cerebellopontine angle in patients with unilateral idiopathic sudden sensorineural hearing loss: Evaluations by three radiological grading systems11
This work aimed to investigate the impact of the position, configuration and neurovascular contact of the anterior inferior cerebellar artery in cerebellopontine angle and internal auditory canal on the clinical features of patients with unilateral idiopathic sudden sensorineural hearing loss (ISSNHL).
- Cross-domain Knowledge Distillation for Retrieval-based Question Answering Systems11
This work provides a new perspective on the potential of the teacher-student paradigm facilitating cross-domain transfer learning, where the teacher and student tasks belong to heterogeneous domains, with the goal to improve the student model’s performance in the target domain.
- 11
- Diffutoon: High-Resolution Editable Toon Shading via Diffusion Models10
An innovative form of toon shading based on diffusion models, aiming to directly render photorealistic videos into anime styles, is proposed called Diffutoon, capable of rendering remarkably detailed, high-resolution, and extended-duration videos in anime style.
- 10
- 10
- Modeled methylmercury exposure and risk from rice consumption for vulnerable populations in a traditional fish-eating area in China10
A food chain model based on the aquivalence method was established to identify major sources of methylmercury (MeHg), estimate the levels of MeHG, and quantify exposure to MeHg via rice and aquatic food consumption, which indicated that pregnant women could face risk from MeH g exposure.
- 9
- Boosting Gradient Leakage Attacks: Data Reconstruction in Realistic FL Settings9
To alleviate the performance bottlenecks identified above, FedLeak is developed, which introduces two novel techniques, partial gradient matching and gradient regularization, and a practical evaluation protocol grounded in a thorough review of extensive FL literature and industry practices is formulated.
- 9
- 9
- 9
- 9
- 9
- Drug design and DNA structural research inspired by the Neidle laboratory: DNA minor groove binding and transcription factor inhibition by thiophene diamidines9
Comparisons of the DNA interactions of diverse diamidine derivatives with a 5’-GAATTC-3’ binding site using a powerful array of methods provide an excellent understanding of how thiophene compounds recognize the minor groove and can act as transcription factor inhibitors.
- 9
- A Novel Approach for Traffic Anomaly Detection in Power Distributed Control System and Substation System9
This paper proposes an novel traffic anomaly detection model for power ICS based on Multi-Head attention (MHA) method, in which the collected raw traffic was converted into the form of matrix, and outperforms the CNN-based and classical machine learning detection models.
- 9
- 9
- 9
- 9
- Efficient Resource Allocation Algorithm for Maximizing Operator Profit in 5G Edge Computing Network8
A novel algorithm to solve the multi-objective resource allocation problem in 5G edge computing (EC) network, the objective is to maximize the operator profit and minimize the total completion time of tasks with priorities from the perspective of service operators under time and workload constraints is presented.
- L3A: Label-Augmented Analytic Adaptation for Multi-Label Class Incremental Learning8
Lab-Augmented Analytic Adaptation (L3A), an exemplar-free approach without storing past samples, which outperforms existing methods in MLCIL tasks and derives a closed-form solution for neural networks.
- CoRec: An Efficient Internet Behavior-based Recommendation Framework with Edge-cloud Collaboration on Deep Convolution Neural Networks8
This work proposes an efficient internet behavior-based recommendation framework with edge-cloud collaboration on deep CNNs (CoRec) to improve both the accuracy and speed for mobile recommendation and introduces a novel convolutional interest network (CIN) that improves the accuracy by modeling the long- and short-term interests and accelerates the prediction through parallel-friendly convolutions.
- 8
- ID Preserving Face Super-Resolution Generative Adversarial Networks8
An ID Preserving Face Super-Resolution Generative Adversarial Networks (IP-FSRGAN) to reconstruct realistic super-resolution face images from low-resolution ones and demonstrates excellent robustness under different downsample scaling factors and extensibility to various face verification models is proposed.
- Intra-Familial Phenotypic Heterogeneity and Telomere Abnormality in von Hippel- Lindau Disease: Implications for Personalized Surveillance Plan and Pathogenesis of VHL-Associated Tumors8
This study provides definitive evidence and possible mechanisms of intra-familial phenotypic heterogeneity in VHL families, which is helpful to the update of surveillance guidelines.
- 8
- 7
- ExVideo: Extending Video Diffusion Models via Parameter-Efficient Post-Tuning7
This work proposes a novel post-tuning methodology for video synthesis models, called ExVideo, designed to enhance the capability of current video synthesis models, allowing them to produce content over extended temporal durations while incurring lower training expenditures.
- HEN: a novel hybrid explainable neural network based framework for robust network intrusion detection7
A hybrid explainable neural network-based framework that improves both the interpretability of the model and the performance in detecting new attacks through the innovative application of the explainable artificial intelligence (XAI) method is proposed.
- aDMSCN7
This work proposes an attention-based Deep Multi-instance Sequential Cross Network (aDMSCN) to deal with the UIP task, and is the first attempt to model the problem from a MIL perspective.
- 7
- Analytic Continual Test-Time Adaptation for Multi-Modality Corruption6
A novel approach, Multi-modality Dynamic Analytic Adapter (MDAA), to tackle MM-CTTA tasks, which introduces analytic learning-a closed-form training technique-through Analytic Classifiers (ACs) to mitigate catastrophic forgetting and the Dynamic Late Fusion Mechanism (DLFM) to dynamically select and integrate reliable information from different modalities.
- 6
- Angular Trajectory of the Vestibular Aqueduct in a Cohort of Chinese Patients with Unilateral Ménière’s Disease: Association with Other Imaging Indices and Clinical Profiles6
ATVA ≥ 140°, an indicator of a hypoplastic endolymphatic sac, was found in approximately one-third of the affected and unaffected ears of patients with MD, as well as in a minority of controls, which suggests that the indices may be a predisposing factor rather than a specific marker for the MD ear.
- FedMCP: Parameter-Efficient Federated Learning with Model-Contrastive Personalization6
FedMCP is proposed, a novel parameter-efficient fine-tuning method with model-contrastive personalization for FL that achieves substantial performance improvements over state-of-the-art FL fine-tuning approaches for PLMs.
- Comparison of vestibular aqueduct visualization on computed tomography and magnetic resonance imaging in patients with Ménière’s disease6
The VA visualization on 3D-SPACE MRI is poorer than that observed on CT and may be affected by its osseous configuration, which may provide a basis for further characterization of VA demonstrated by MRI and its clinical significance.
- XtremeCLIP: Extremely Parameter-efficient Tuning for Low-resource Vision Language Understanding6
In this paper, a simple yet efficient paradigm for low-resource VLU named XtremeCLIP is introduced, which involves very few trainable parameters to improve the generalization ability of the trained models.
- 6
- 6
- 6
- Decompose, Plan in Parallel, and Merge: A Novel Paradigm for Large Language Models based Planning with Multiple Constraints5
A novel parallel planning paradigm, which decomposes the complex task based on constraints into subtasks, generates the subplan for each subtask in parallel, and merges them into a global plan (DPPM).
- Development of universal influenza vaccines: strategies for broadly cross-reactive influenza vaccine responses5
This review highlights key advances and persistent challenges in the development of broadly cross-reactive influenza vaccines, with emphasis on conserved antigen design, mucosal immunity, delivery strategies, and future directions toward achieving universal protection against both seasonal and pandemic influenza strains.
- Disseminated tumor cells in bone marrow as predictive classifiers for small cell lung cancer patients5
Bone marrow sampling and characterization of DTC subtypes provided a valuable tool for predicting treatment response and the prognosis in SCLC, and indicated that DTCs inherit a greater amount of homologous somatic information from pre-treatment CTCs, indicating their potential role in disease progression and treatment resistance.
- Joining Natural and Synthetic DNA Using Biversal Nucleotides: Efficient Sequencing of Six-Nucleotide DNA5
This work introduces a new way to connect synthetic and standard molecular biology: biversal nucleotides, which can enable communication across and between standard DNA and synthetic DNA more generally.
- 5
- SueNes: A Weakly Supervised Approach to Evaluating Single-Document Summarization via Negative Sampling5
This paper presents a proof-of-concept study to a weakly supervised summary evaluation approach without the presence of reference summaries, and shows a great advantage in gauging linguistic qualities over all metrics.
- 5
- 5
- 5
- 4
- DocAsRef: An Empirical Study on Repurposing Reference-based Summary Quality Metrics as Reference-free Metrics4
After being repurposed reference-freely, the zero-shot BERTScore using the pretrained DeBERTa-large-MNLI model of<0.5B parameters consistently outperforms its original reference-based version across various aspects on the SummEval and Newsroom datasets.
- FastBlend: a Powerful Model-Free Toolkit Making Video Stylization Easier4
A powerful model-free toolkit called FastBlend is proposed to address the consistency problem for video processing, based on a patch matching algorithm, and it outperforms existing methods for video deflickering and video synthesis.
- 4
- Transferable Adversarial Examples with Bayes Approach4
By regarding the transferability of AEs as generalization ability of the model, it is revealed that vanilla black-box attacks craft AEs via solving a maximum likelihood estimation (MLE) problem, which is an effective approach to boost the generalization of results with limited available data.
- 4
- Defense against Backdoor Attacks via Identifying and Purifying Bad Neurons4
A novel backdoor defense method to mark and purify the infected neurons in the backdoored neural networks, called benign salience, and a new Adaptive Regularization (AR) mechanism is proposed to assist in purifying these identified infected neurons via fine-tuning.
- G-CNN and double-referenced thresholding for detecting time series anomalies4
A so-called G-CNN architecture that combined the gated recurrent units (GRU) with a convolutional neural network (CNN) is proposed, which respectively focus on the monitored and exogenous data.
- 4
- 4
- 4
- 4
- 4
- 4
- 3
- 3
- Log anomaly detection method based on CNN and LSTM fusion3
A method named CLSTMlog is introduced, which combines Convolutional Neural Network (CNN) and Long Short-Term Memory (LSTM) networks, making the most of CNN's ability to extract local features and LSTM's capability to handle temporal relationships.
- 3
- Enhancing Portfolio Optimization via Heuristic-Guided Inverse Reinforcement Learning with Multi-Objective Reward and Graph-based Policy Learning3
This work proposes a novel heuristic-guided inverse reinforcement learning framework for portfolio optimization that provides an interpretable expert strategy generation mechanism that takes into account sector diversification and correlation constraints and utilizes heterogeneous graph policy learning with hierarchical attention mechanisms to explicitly model inter-stock relationships.
- DualToken-ViT: Position-aware Efficient Vision Transformer with Dual Token Fusion3
This paper proposes a light-weight and efficient vision transformer model called DualToken-ViT that leverages the advantages of CNNs and ViTs and effectively fuses thetoken with local information obtained by convolution-based structure and the token with global information obtained from self-attention-based structures to achieve an efficient attention structure.
- Rapid Diffusion: Building Domain-Specific Text-to-Image Synthesizers with Fast Inference Speed3
This paper proposes Rapid Diffusion, a novel framework for training and deploying super-resolution, text-to-image latent diffusion models with rich entity knowledge injected and optimized networks and employs BladeDISC and FlashAttention techniques to optimize computational graphs of the generated models for online deployment.
- ParaSum: Contrastive Paraphrasing for Low-Resource Extractive Text Summarization3
This paper proposes a novel paradigm for low-resource extractive summarization, called ParaSum, which reformulates text summarization as textual paraphrasing, and minimizes the training gap between the summarization model and PLMs, enabling a more effective probing of the knowledge encoded within PLMs and enhancing the summarizations performance.
- 3
- 3
- Refiner: Data Refining against Gradient Leakage Attacks in Federated Learning3
A novel defensive paradigm that departs from conventional gradient perturbation approaches and instead focuses on the construction of robust data is explored, and empirical evaluations demonstrate the superior defense effectiveness of Refiner at defending against state-of-the-art attacks.
- 3
- A large-scale task scheduling algorithm based on clustering and duplication3
A task scheduling algorithm that can decrease the scheduling time of large-scale tasks on a limited number of processors and speed up the global execution time of the task is proposed.
- 3
- 3
- Understand the Effect of Mud Pulse on Drilling Dynamics Using Big Data and Numerical Modeling3
Large mud pressure pulse utilized by some MWD tools introduces significant rpm variation and shock and vibration, which is quantified by big data and further demonstrated by drilling system modeling.
- Street Sign Recognition Algorithm Based on Deep Learning3
This paper improved the existing object detection algorithm SSD, and made it possible to detect text objects in traffic guidance sign using a deep neural network CRNN to identify the text.
- 3
- Research and Realization of Commodity Image Retrieval System Based on Deep Learning3
A commodity image retrieval system based on CNN and ListNet sort learning method that combines the learn to rank algorithm of the text retrieval domain and stochastic gradient descent algorithm, solving the problem of the image retrieval.
- Multi-Domain Fragmentation-Aware RSA Operations through Cooperative Hierarchical Controllers in SD-EONs3
This paper investigates lightpath provisioning with fragmentation-aware routing and spectrum assignment across multiple software-defined elastic optical network (SD-EON) domains, with a hierarchical controller framework.
- ConDiffFuzz: Dependency-Aware Consistency Checking for Differential Fuzzing of Industrial Control Protocol Implementations2
ConDiffFuzz is proposed, a dependency-aware and dynamically adjusted hierarchical consistency checking method for industrial control protocol implementations that analyzes dependencies among check sequences to optimize execution order and dynamically prunes and regenerates dependent sequences after failures to mitigate inconsistency error propagation.
- Pre-treatment Circulating Tumor Cell Associated White Blood Cell Clusters Independently Predict Poor Survival in Patients with Extensive-disease Small Cell Lung Cancer2
The heterogeneity and prognostic significance of various aneuploid circulating tumor cells (CTCs) subtypes and CTC-associated white blood cell (CTC-WBC) clusters in patients with LD-and ED-SCLC respectively are investigated.
- 2
- TOTF: Missing-Aware Encoders for Clustering on Multi-View Incomplete Attributed Graphs2
This work investigates the problem of multi-view incomplete attributed graph clustering for the first time, and proposes a correlation strength-aware graph neural network on the basis of the inherent relationships among attributes to enhance accuracy.
- Maximizing the number of completed tasks in MEC considering time and energy constraints2
The mayfly genetic algorithm (MGA) is developed, which jointly optimizes task offloading locations and ratios, central processor unit (CPU) frequencies of user devices and computing capacities allocated to user devices by edge servers and indicates that MGA outperforms state-of-the-art algorithms in terms of the number of completed tasks.
- 2
- 2
- 2
- 2
- A Python script to merge Sanger sequences2
This study provides a Python script that is able to assemble multiple overlapping Sanger sequences and offers a simple and free method for merging Sangers sequences and is useful for gene cloning.
- 2
- End-to-end Semantics-based Summary Quality Assessment for Single-document Summarization.2
This work introduces a new end-to-end metric system for summary quality assessment by leveraging the semantic similarities of words and/or sentences in deep learning and exhibits very promising results on gold-standard datasets and suggests its great potential to future summarization research.
- Research on UAV Image Classification Based on Deep Learning2
The main research result of this algorithm is to propose a key frame extraction scheme for UAV image and then combine it with the application of Mask R-CNN in high-altitude image to identify the ground area.
- Fast robust matching algorithm based on BRISK and GMS2
An improved image registration algorithm based on BRISK and GMS that achieves higher matching accuracy on the basis of maintaining speed compared with the original algorithm, and the average accuracy is improved by 8.02%.
- 2
- 2
- 1
- CCJA: Context-Coherent Jailbreak Attack for Aligned Large Language Models1
This work highlights the urgency of stronger safety alignment and provides a practical framework for worst-case robustness evaluation of open-source models by developing an efficient and interpretable attack method that minimizes manual intervention.
- Efficient Inference Scheduling With Edge and Cloud Collaboration for LLMs Under Resource Constraints1
- GSDNet: Revisiting Incomplete Multimodal-Diffusion from Graph Spectrum Perspective for Conversation Emotion Recognition1
A novel graph diffusion network, GSDNet, which maps Gaussian noise to the graph spectral space of missing modalities and recovers the missing data according to its original distribution, and only affects the eigenvalues of the adjacency matrix instead of destroying the adjacency matrix directly.
- 1
- 1
- 1
- Research on distributed PV network terminal identification technology based on device fingerprint generation1
An intelligent device fingerprint access technology for distributed photovoltaic power network system based on adaptive filtering, which can realize the device authentication of the distributed photovoltaic network intelligent terminal equipment access, so as to effectively improve the comprehensive defense capability of the distributed photovoltaic network.
- A modeling method for cyber-physical attacks on distributed photovoltaic systems based on probabilistic colored Petri nets1
A modeling method based on Probabilistic Colored Logical Petri Nets (PCLPN) to describe and analyze cyber-physical attacks in DPS is proposed and the effectiveness of the proposed method is verified in capturing the complex dynamic behaviors of DPS and evaluating system vulnerabilities.
- Sentiment Confidence Separation: A Trust-Optimized Framework for Multimodal Sentiment Classification1
- Network security enhancement strategies for power distribution station based on attack graphs1
The experimental results show that the proposed scheme can find the key attack nodes more efficiently in the network attack graph of the power distribution station.
- 1
- 1
- Development of Multi-modal Physiological Signals Acquisition Platform and Its Application in Emotional Recognition Based on a Small Number of Samples1
The experimental results show that the multi-modal physiological signals acquisition platform built in this paper is effective and the emotions of subject can be recognized in specific scenes based on a small number of samples.
- Determinantal point process-based new radio unlicensed link scheduling for multi-access edge computing1
A determinantal point process (DPP) based link scheduling to improve the terminals downlink performance in the hidden nodes zone and demonstrates that the proposed DPP method is capable of effectively selecting the optimal transmitting nodes and optimizing more hidden nodes while maximizing the sum rate.
- 1
- 1
- 1
- 1
- 1
- 1
- 1
- 1
- Automatic Knowledge Fusion in Transferrable Networks for Semantic Text Matching1
This work proposes a novel method to automatically fuse transferable network architecture that has better performance than the state-of-the-art transfer learning architectures.
- 1
- 1
- 1
- Bone Age Assessment based on Finger joint Localization1
A finger joint localization strategy to filter out most non-informative parts of images in the forecast of skeletal bone age with the specifically combine joints images to increase the performance accuracy compared with the whole hand images.
- 1
- 1
- 1
- 1
- –
- –
- Brain‐Inspired Memory Network Empowering Intelligent Power System Inspection–
A robust tracking method for intelligent power system inspections is introduced, leveraging the brain‐inspired memory network (BIMN) to enhance the accuracy and reliability of target tracking in complex and dynamic environments.
- PointShuffler: Accelerating Point Cloud Neural Networks on General-Purpose GPUs–
PointShuffler, the first-of-its-kind GPU-friendly inference engine that efficiently eliminates PCNN redundancy with preserved accuracy is proposed, which consists of three enabling optimizations.
- S-GPT: Shape-aware graph partitioning and tuning for sustainable inference of large-scale deep neural networks–
S-GPT is proposed, a shape-aware compiler framework that co-optimizes graph-level partitioning and backend kernel tuning and develops a shape-aware kernel tuning mechanism, where structurally similar subgraphs and neighboring shape intervals can reuse historical schedules, measurement records, and cost-model states.
- Cross-Modality Anatomy Knowledge Distillation for Retinal OCT Layer Segmentation With Limited Labels–
- –
- –
- –
- –
- Abstract 6076: Non-clinical pharmacology of HX044, a novel FIC CTLA4xCD47 bi-specific antibody–
The anti-tumor efficacy of HX044 is significantly superior to Ipilimumab or SIRPα-Fc fusion protein at equivalent molar doses, respectively, and preclinical data of HX044 support its further clinical testing in cancer patients.
- Detection and diagnosis of unknown threats in power equipment using machine learning and Spark technology–
This research utilizes a lightweight gradient-based method for detecting known threats, and proposes a novel detection approach for unknown threats that combines classical anomaly detection methods, specifically support vector machines, with autoencoders.
- –
- –
- Neurovascular compression and vestibular compromises in idiopathic sudden sensorineural hearing loss–
V Vestibular testing results and radiological neurovascular compression signs in CPA-IAC region may not be associated in ISSNHL patients, and vestibular impairment is more closely related to hearing outcomes than neurovascular compression.
- –
- –
- –
- –
- –
- –
- F-OAL: Forward-only Online Analytic Learning with Fast Training and Low Memory Footprint in Class Incremental Learning–
Instead of back-propagation, the Analytic Classifier (AC) updated by recursive least square, cooperating with a frozen backbone is designed, which simultaneously achieves high accuracy, low resource consumption and data privacy protection.
- –
- –
- TransPrompt v2: A Transferable Prompting Framework for Cross-task Text Classification–
Extensive experiments show that TransPrompt v2 outperforms single-task and cross-task strong baselines over multiple NLP tasks and datasets, and shows that the meta-learner can effectively improve the performance of PLMs on previously unseen tasks.
- Mimicry API Gateway Decision Algorithm Based on Trust Distribution–
An adjudication algorithm for mimic API gateway based on trust distribution comprehensively considers the direct and indirect trust relationships between equipment executives from the perspective of trust distribution, and introduces a dynamic penalty strategy, which greatly improves the dynamics and effectiveness of the algorithm.
- –
- –
- –
- –
- Abstract 6345: Non-clinical pharmacology of HX009, a novel FIC PD1xCD47 BsAb–
A novel rationally designed bi-specific antibody (BsAb) targeting both PD1 and CD47, HX009, is reported in order to improve both efficacy and safety over the two respective single-targeting ICIs.
- Abstract 5665: HX301 (ON1232580) a novel kinase inhibitor with potent activity against CSF1R and FLT3, shows strong anti-AML activity in defined preclinical models–
The data suggests that HX301 can potentially be explored for the treatment of AML, at least a subset of the patients with CSF1R and FLT3-ITD as leukemogenic drivers, and believes that Hx301 could be a potential candidate for treating subset of AMl, warranting further clinical investigation.
- Software family detection based on behavior analysis–
The role of features in malicious behavior identification and the characteristics of various detection methods, which can help security researchers to break through the limitations of traditional detection technology are clarified.
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- Nucleic acid crystallization and phase determination facilitated by selenium functionalization–
This paper presents the most recent selenium-atom functionalization of nucleic acids and their potential applications in 3D structure-and-function studies and anticancer therapeutics in molecular medicine, and will provide important tools to investigate nucleic acid structure/folding, recognition and catalysis.
- –
- Immobilizing hydroxycholesterol with apatite on titanium surfaces to induce ossification–
In this study, hydroxycholesterol was immobilized with apatite on titanium through simply adsorption or sandwich-like coating, which induced the ossification in vitro and in vivo.
- –
- –
- –
- –
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 2026-10-12. 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.