A5100697510- Cloud Computing and Resource Management
- Topic Modeling
- Stochastic Gradient Optimization Techniques
- Parallel Computing and Optimization Techniques
- Privacy-Preserving Technologies in Data
- Machine Learning and ELM
- Interconnection Networks and Systems
- Natural Language Processing Techniques
- Neural Networks and Applications
- Higher Education and Teaching Methods
- Optical Network Technologies
- Machine Learning and Data Classification
- Machine Learning and Algorithms
- Face and Expression Recognition
- Mobile Ad Hoc Networks
- Opportunistic and Delay-Tolerant Networks
- Advanced Photonic Communication Systems
- Advanced Neural Network Applications
- Advanced Computational Techniques and Applications
- Data Stream Mining Techniques
- VLSI and Analog Circuit Testing
- Education and Work Dynamics
- Domain Adaptation and Few-Shot Learning
- Technology and Data Analysis
- Mobile Crowdsensing and Crowdsourcing
Nanjing University of Information Science and Technology
2025
Beijing University of Posts and Telecommunications
2025
Qilu University of Technology
2024
Shandong Academy of Sciences
2024
Beijing Union University
2022-2024
Jilin Electric Power Research Institute (China)
2024
Rice University
2021-2023
Central University of Finance and Economics
2023
Dalian Maritime University
2023
China University of Geosciences
2023
Success of manufacturing companies largely depends on reliability their products. Scheduled maintenance is widely used to ensure that equipment operating correctly so as avoid unexpected breakdowns. Such often carried out separately for every component, based its usage or simply some fixed schedule. However, scheduled labor-intensive and ineffective in identifying problems develop between technician's visits. Unforeseen failures still frequently occur. In contrast, predictive techniques help...
Abstract In silico methods are increasingly important in predicting the ecotoxicity of engineered nanomaterials (ENMs), encompassing both individual and mixture toxicity predictions. It is widely recognized that ENMs trigger oxidative stress effects by generating intracellular reactive oxygen species (ROS), serving as a key mechanism their cytotoxicity studies. However, existing still face significant challenges induced ENMs. Herein, we utilized laboratory-derived data machine learning to...
Large deep learning models have recently garnered substantial attention from both academia and industry. Nonetheless, frequent failures are observed during large model training due to large-scale resources involved extended time. Existing solutions significant failure recovery costs the severe restriction imposed by bandwidth of remote storage in which they store checkpoints.
Kernel support vector machines (SVMs) deliver state-of-the-art results in many real-world nonlinear classification problems, but the computational cost can be quite demanding order to maintain a large number of vectors. Linear SVM, on other hand, is highly scalable data only suited for linearly separable problems. In this paper, we propose novel approach called low-rank linearized SVM scale up kernel limited resources. Our transforms linear one via an approximate empirical map computed from...
Gradient compression (GC) is a promising approach to addressing the communication bottleneck in distributed deep learning (DDL). It saves time, but also incurs additional computation overheads. The training throughput of compression-enabled DDL determined by strategy, including whether compress each tensor, type compute resources (e.g., CPUs or GPUs) for compression, schemes compressed and so on. However, it challenging find optimal strategy applying GC because intricate interactions among...
This paper describes a Security Enhanced AODV routing protocol for wireless mesh networks (SEAODV). SEAODV employs Blom's key pre-distribution scheme to compute the pairwise transient (PTK) through flooding of enhanced HELLO message and subsequently uses established PTK distribute group (GTK). GTK are used authenticating unicast broadcast messages respectively. In networks, unique is shared by each pair nodes, while secretly between node all its one-hop neighbors. A authentication code (MAC)...
Abstract Carriage belt conveyors represent a novel type of bulk material handling system with distinctive wheel-rail contact features. The turnover mechanism enables the carriages to transition seamlessly between carry and return segments, facilitating continuous cyclic operation. When wire rope interacts pulley, presence disrupts full wheel, resulting in dynamic polygonal structure. However, existing research has given limited attention effect rope. To address this gap, study proposes...
Double-endobutton technique, as a widely accepted strategy for the treatment of acromioclavicular joint dislocation, is undergoing constant improvement. This study aims to assess clinical effect modified single-endobutton combined with nice knot in fixation Rockwood type III or V dislocation.From January 2016 June 2019, 16 adult patients (13 males and 3 females) dislocation were treated technique our department. The age ranged from 18 64 years old an average 32.8 old. Operative time,...
Many viruses use the host cell division cycle to facilitate replication. Cyclin-dependent kinases (CDKs) are a group of serine/threonine that play central role in regulating progression. However, prospect using CDKs for anti-influenza virus treatment remains be elucidated. We conducted this study investigate potential CDK1 inhibitor Ro-3306 preventing influenza infection and elucidate underlying mechanism. showed Ro-3306, inhibitor, exerts activity both vitro vivo. Proof-of-concept studies...
The Zika virus (ZIKV) has garnered significant public attention, particularly following the outbreak in Brazil, due to its potential cause severe damage central nervous system and ability cross placental barrier, resulting microcephaly infants. Despite urgency, there remains a lack of targeted therapies or vaccines for prevention treatment ZIKV infection related diseases. Fangchinoline (FAN), an alkaloid derived from traditional Chinese medicinal herbs, range biological activities. In this...
Kernel perceptrons are represented by a subset of training points, called the support vectors, and their associated weights. To address issue unlimited growth in model size during training, budget kernel maintain fixed number vectors thus achieve constant update time space complexity. In this paper, new perceptron algorithm for online learning on is proposed. Following idea tighter perceptron, upon exceeding budget, removes vector with minimal impact classification accuracy. optimize memory...
A fast online algorithm OnlineSVM <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">R</sup> for training Ramp-Loss Support Vector Machines (SVM s) is proposed. It finds the optimal SVM t + 1 examples using SVMR built on previous examples. The retains Karush-Kuhn-Tucker conditions all previously observed This achieved by an SMO-style incremental learning and decremental unlearning under Concave-Convex Procedure framework. Further speedup of time...
Traffic aggregates in cloud data center networks are by and large buffered transmitted simple physical FIFO queues. Despite the crucial role they play, a well-known problem of queues is that unable to provide precise bandwidth guarantees. This leads range negative impacts spanning application layer, transport link layer.