Zejun Jiang

ORCID: 0000-0002-6651-0826
Publications
Citations
Views
---
Saved
---
About
Contact & Profiles
Research Areas
  • Advanced Computational Techniques and Applications
  • Topic Modeling
  • Natural Language Processing Techniques
  • Cloud Computing and Resource Management
  • Service-Oriented Architecture and Web Services
  • Caching and Content Delivery
  • Advanced Data Storage Technologies
  • Embedded Systems and FPGA Design
  • Speech and dialogue systems
  • Power Systems and Technologies
  • Multimodal Machine Learning Applications
  • Distributed and Parallel Computing Systems
  • Web Applications and Data Management
  • Advanced Graph Neural Networks
  • Peer-to-Peer Network Technologies
  • Industrial Technology and Control Systems
  • Cloud Data Security Solutions
  • Opportunistic and Delay-Tolerant Networks
  • Real-Time Systems Scheduling
  • Parallel Computing and Optimization Techniques
  • Advanced Battery Technologies Research
  • Fuel Cells and Related Materials
  • Pesticide Residue Analysis and Safety
  • Electrocatalysts for Energy Conversion
  • Privacy-Preserving Technologies in Data

Jilin Electric Power Research Institute (China)
2024

Northwestern Polytechnical University
2013-2024

Harbin Engineering University
2023-2024

Guangdong University of Technology
2022

China Jiliang University
2022

Suzhou University of Science and Technology
2010

Tianjin University
2008

A kind of hybrid fiber interferometer consisting a Sagnac (FSI), closed-cavity Fabry-Perot (FPI), and an open-cavity FPI is proposed for generating combined-Vernier-effect. Through adjusting the polarization-maintaining (PMF) length FSI, free spectral range (FSR) tailored to be similar that parallel-connected reference producing first Vernier effect, which spectrum used match sensing obtaining second effect. Noticeable lower upper envelopes are achieved in effects, respectively, so called...

10.1364/oe.454988 article EN cc-by Optics Express 2022-03-04

Abstract Cloud Storage services are increasingly noticed as they promise elastic capability and high reliability at low cost. In such services, you can store most of your files to authenticated Service center, do not worry about space being inadequate or wasted because the storage able be adjusted dynamically is important feature Storage. this paper, we present a solution how build System based on open-source distributed database, it follows stratum design that includes Web service...

10.1016/j.proenv.2011.09.111 article EN Procedia Environmental Sciences 2011-01-01

Three-dimensional flow calculations of the in-cylinder for a direct-injection Diesel engine have been carried out with different combustion chambers. In order to assess decisive factors formation and evolution field, simplified piston bowls were used. A limited number validation compression stroke performed. The results show that geometries significant influence on distribution cross section-averaged swirl ratio. tangential velocity profiles at sections are highly nonlinear, rigid body...

10.1080/19942060.2008.11015235 article EN Engineering Applications of Computational Fluid Mechanics 2008-01-01

Multi-hop machine reading comprehension (MRC) task aims to enable models answer the compound question according bridging information. Existing methods that use graph neural networks represent multiple granularities such as entities and sentences in documents update all nodes synchronously, ignoring fact multi-hop reasoning has a certain logical order across granular levels. In this paper, we introduce an Asynchronous Multi-grained Graph Network (AMGN) for MRC. First, construct multigrained...

10.24963/ijcai.2021/531 article EN 2021-08-01

HIV-1 protease cleavage site is critical for the design of inhibitors. Classification algorithms based on traditional machine learning are often used to deal with prediction sites. Unlike classification learning, deep can extract characteristics data well and get better performance. In this paper, innovatively converted One-hot data, then two models proposed RNN LSTM. At last, experimental results compared support vector algorithm random forest in algorithm. The show that network structure...

10.1109/icsai.2018.8599496 article EN 2018-11-01

Conversational emotion recognition (CER) is a significant task due to its application in human–computer interaction. Existing work treats CER as an utterance-level classification without considering that empathic response also reflects contextual understanding. Previous has proven accurate of emotions the dialogue history helpful generate high-fit responses. In this paper, we investigate whether conclusion sufficient and necessary condition. Specifically, define auxiliary multiturn...

10.1002/int.22805 article EN International Journal of Intelligent Systems 2022-01-06

Many real-time analytical applications over massive data streams were performed by usually introducing a specific stream processing core. In general, these SPCs not popularly applied to enterprises same as MapReduce, even if now analytics are taken into attention more and more. For reversing this tide, we developed new system. Our system modified the stock Hadoop's MapReduce programming model execution framework, used Chord temporary data, Cassandra its persistent storage. With our system,...

10.1109/icmlc.2012.6359554 article EN International Conference on Machine Learning and Cybernetics 2012-07-01

An energy management strategy (EMS) is important in marine hybrid propulsion systems while the global optimal solutions are more significant for EMS to improve fuel economy of system. Aiming at finding accurate EMS, an improved dynamic programing (DP) algorithm was proposed overcome basic DP algorithms’ weaknesses and economy. Then model gas-electric system also established verify strategies. What’s more, applied minimum natural gas consumption charge-sustaining strategy. And a bi-level...

10.1177/14680874231195279 article EN International Journal of Engine Research 2023-09-12

To maximize the writing throughput of deduplication system, most systems and clusters sequentially store new chunks in disk. This method results data fragments as system grows. It is important to analyse understand its features. We features using three datasets from real world. utilize File Fragment Degree (FFD) quantize a file system. firstly implement Extreme Binning (EB) collect chunk addresses every dataset. Then, we design FFD analyser compute for according sizes. Finally, numbers. As...

10.1109/icsse.2012.6257249 article EN 2012-06-01
Coming Soon ...