Search

Your search keyword '"Xia, Chengyu"' showing total 161 results
Start Over Author "Xia, Chengyu"
161 results on '"Xia, Chengyu"'

1. Bayesian Federated Model Compression for Communication and Computation Efficiency

Author

Xia, Chengyu, Tsang, Danny H. K., and Lau, Vincent K. N.

Subjects
Computer Science - Machine Learning, Computer Science - Artificial Intelligence, Computer Science - Distributed, Parallel, and Cluster Computing
Abstract

In this paper, we investigate Bayesian model compression in federated learning (FL) to construct sparse models that can achieve both communication and computation efficiencies. We propose a decentralized Turbo variational Bayesian inference (D-Turbo-VBI) FL framework where we firstly propose a hierarchical sparse prior to promote a clustered sparse structure in the weight matrix. Then, by carefully integrating message passing and VBI with a decentralized turbo framework, we propose the D-Turbo-VBI algorithm which can (i) reduce both upstream and downstream communication overhead during federated training, and (ii) reduce the computational complexity during local inference. Additionally, we establish the convergence property for thr proposed D-Turbo-VBI algorithm. Simulation results show the significant gain of our proposed algorithm over the baselines in reducing communication overhead during federated training and computational complexity of final model.

Published
2024

5. Structured Bayesian Compression for Deep Neural Networks Based on The Turbo-VBI Approach

Author

Xia, Chengyu, Tsang, Danny H. K., and Lau, Vincent K. N.

Subjects
Computer Science - Machine Learning
Abstract

With the growth of neural network size, model compression has attracted increasing interest in recent research. As one of the most common techniques, pruning has been studied for a long time. By exploiting the structured sparsity of the neural network, existing methods can prune neurons instead of individual weights. However, in most existing pruning methods, surviving neurons are randomly connected in the neural network without any structure, and the non-zero weights within each neuron are also randomly distributed. Such irregular sparse structure can cause very high control overhead and irregular memory access for the hardware and even increase the neural network computational complexity. In this paper, we propose a three-layer hierarchical prior to promote a more regular sparse structure during pruning. The proposed three-layer hierarchical prior can achieve per-neuron weight-level structured sparsity and neuron-level structured sparsity. We derive an efficient Turbo-variational Bayesian inferencing (Turbo-VBI) algorithm to solve the resulting model compression problem with the proposed prior. The proposed Turbo-VBI algorithm has low complexity and can support more general priors than existing model compression algorithms. Simulation results show that our proposed algorithm can promote a more regular structure in the pruned neural networks while achieving even better performance in terms of compression rate and inferencing accuracy compared with the baselines.

Published
2023
Full Text
View/download PDF

18. Structured Bayesian Federated Learning for Green AI: A Decentralized Model Compression Using Turbo-VBI Based Approach

Author

Xia, Chengyu, Tsang, Hin Kwok, Lau, Kin Nang, Xia, Chengyu, Tsang, Hin Kwok, and Lau, Kin Nang

Abstract

Although deep neural networks (DNNs) have been remarkably successful in numerous areas, the performance of DNN is compromised in federated learning scenarios because of the large model size. A large model can induce huge communication overhead during the federated training, and also induce infeasible storage and computation burden at the clients during the inference. To address these issues, we investigate structured model compression in federated learning to construct sparse models with regular structure such that they require significantly less communication, storage and computation resources. We do this by proposing a three-layer hierarchical prior, which can promote a common regular sparse structure in the local models. We design a decentralized Turbo variational Bayesian inference (D-Turbo-VBI) algorithm to solve the resulting federated training problem. With the common regular sparse structure, both upstream and downstream communication overhead can be reduced, and the final model also has a regular sparse structure, which requires significantly less local storage and computation resources. Simulation results demonstrate that our proposed algorithm can efficiently reduce the communication overhead during federated training and the resulting model can achieve a significantly lower sparsity rate and inference time compared to the baselines while maintaining a competitive accuracy. IEEE

Published
2024

19. Multi-Resolution Model Compression for Deep Neural Networks: A Variational Bayesian Approach

Author

Xia, Chengyu, Guo, Huayan, Ma, Haoyu, Tsang, Hin Kwok, Lau, Kin Nang, Xia, Chengyu, Guo, Huayan, Ma, Haoyu, Tsang, Hin Kwok, and Lau, Kin Nang

Abstract

The continuously growing size of deep neural networks (DNNs) has sparked a surge in research on model compression techniques. Among these techniques, multi-resolution model compression has emerged as a promising approach which can generate multiple DNN models with shared weights and different computational complexity (resolution) through a single training. However, in most existing multi-resolution compression methods, the model structures for different resolutions are either predefined or uniformly controlled. This can lead to performance degradation as they fail to implement systematic compression to achieve the optimal model for each resolution. In this paper, we propose to perform multi-resolution compression from a Bayesian perspective. We design a resolution-aware likelihood and a two-layer prior for the channel masks, which allow joint optimization of the shared weights and the model structure of each resolution. To solve the resulted Bayesian inference problem, we develop a low complexity partial update block variational Bayesian inference (PUB-VBI) algorithm. Furthermore, we extend our proposed method into the arbitrary resolution case by proposing an auxiliary neural network (NN) to learn the mapping from the input resolution to the corresponding channel masks. Simulation results show that our proposed method can outperform the baselines on various NN models and datasets.

Published
2024

25. THE LIFE OPTIMIZATION OF THE SEALING SURFACE OF THE ROTATING CONTROL HEAD RUBBER CORE BASED ON THE EXPERIMENT

Author

HE YuHang, HOU ZuoFu, ZHAO JingYing, XIA ChengYu, WU JiWei, HUANG Jian, and FANG Yong

Subjects
Underbalanced drilling, Rotating blowout preventer（RBOP）, Finite element simulation, Experiment and field application, Mechanical engineering and machinery, TJ1-1570, Materials of engineering and construction. Mechanics of materials, TA401-492
Abstract

In the process of underbalance drilling,the rubber core of rotating control head is subject to the alternating cycle load,which is prone to fatigue failure.Firstly,the sealing principle of rotating control head is analyzed and the stress of rubber core in dynamic sealing process is obtained.Then,the Yeoh constitutive model in the deformation process was established by the tensile test of the rubber core material. Then ABAQUS software was used to establish the finite element model of the dynamic sealing of the rubber core,and the simulation of the sealing process of the rubber core during the drilling and the optimization analysis of the factors affecting the fatigue life of the sealing surface of the rubber core were carried out.In this paper,the inner diameter of rubber core is optimized,and the inner diameter size of rubber core should be matched with different outer diameter drill pipes is optimized.The paper also optimizes the lowering and lifting speed of drill pipe under different wellhead pressures.The higher the well pressure,the more likely the rubber core fatigue failure will be. The safe lowering and lifting speed of drill pipe under different pressures is obtained.At last,according to the Chinese standard,indoor seal test and field application test were carried out in the well control equipment quality supervision and inspection center of China national petroleum industry to confirm the reliability of finite element optimization,which is of great practical significance to improve the service life of rubber core and oilfield operation safety.

Published
2020
Full Text
View/download PDF

26. Adult Intramedullary Pilocytic Astrocytomas: Clinical Features, Management, and Outcomes

Author

Yang, Tao, Wu, HaiBo, and Xia, ChengYu

Subjects
Astrocytoma -- Diagnosis -- Care and treatment -- Patient outcomes, Magnetic resonance imaging -- Usage -- Methods, Surgery -- Methods -- Patient outcomes, Health
Abstract

Byline: Tao. Yang, HaiBo. Wu, ChengYu. Xia Purpose: Adult intramedullary pilocytic astrocytomas (PAs) are exceedingly rare. The aim of this study was to summarize our experiences in treating adult intramedullary [...]

Published
2022

37. Research on Cuttings Carrying Principle of New Aluminum Alloy Drill Pipe and Numerical Simulation Analysis.

Author

Wu, Pengcheng, Li, Chentao, Zhang, Zhen, Yang, Jingwei, Gao, Yanzhe, Wang, Xianbing, Wan, Xiumei, Xia, Chengyu, and Guo, Qunying

Subjects
DRILL pipe, ALUMINUM alloys, STEEL pipe, NUMERICAL analysis, COMPUTER simulation, SHALE gas
Abstract

In order to improve the cuttings transport ability and well hole purification effect of horizontal shale gas wells in the Sichuan and Chongqing area, a new type of aluminum alloy drill pipe is put forward, and the floatability is validated by theoretical analysis and actual parameter calculation. According to the cuttings migration mechanism, the cuttings cleaning simulation model of the new aluminum alloy drill pipe and the traditional steel drill pipe was established by Fluent, and the hexahedral mesh is used to divide the model. Thus, the mesh independence and convergence of the two models are verified. Then, the simulation model is verified by comparing the calculation results of the two simulation models with the experimental data of the indoor cuttings migration device. Finally, the rock-cleaning ability of the two drill pipes is analyzed under the conditions of changing the cuttings particle size, well inclination angle, displacement, and mechanical speed. Compared with the traditional steel drill pipe, the new aluminum alloy drill pipe can improve the borehole purification capacity by 13% on average. This research result is of great significance in reducing the quality of cuttings in the annulus and improving the borehole purification effect. [ABSTRACT FROM AUTHOR]

Published
2023
Full Text
View/download PDF

38. Transient prediction of annular pressure between packers in high-pressure low-permeability wells during high-rate, staged acid jobs

Author

Ding Liangliang, Rao Jiyang, and Xia Chengyu

Subjects
Chemical technology, TP1-1185, Energy industries. Energy policy. Fuel trade, HD9502-9502.5
Abstract

For high-pressure low-permeability wells, wellbore temperature drops drastically in high-rate and multistage acid fracturing process. Under the combined action of the swelling of tubing string and the contraction of annular fluid between packers, annular pressure between packers undergoes violent transient change in staged acid jobs, thereby deteriorating loading on the tubing string and packers. Based on the principle of energy conservation and wellbore heat conduction, the transient prediction of two-dimensional (2D) wellbore temperature field under pumping injection condition was established by considering the effects of heat generated by friction and convection heat exchange. Moreover, the effects of wellbore temperature/pressure changes on the annular volume between packers were analyzed. Furthermore, in combination with the transient prediction model of wellbore temperature, PVT state equation of annular fluid, the calculation model of tubing string radial deformation and the transient seepage equation of the formation, the transient prediction model of annular pressure between packers in high-pressure low-permeability wells was established. Finally, by taking a high-pressure low-permeability well as an example, annular pressure between packers was calculated and the forces on the packers and tubing string were analyzed. According to the prediction results, the tubing string, which was regarded to be safe using conventional design method, exhibited an extremely high risk of failure after taking into account the decrease in annular pressure between packers. Therefore, the decrease in annular pressure should be fully considered in the design of tubing string for high-pressure low-permeability wells in multistage acid fracturing process. In combination with sensitivity analysis results, it can be concluded that formation permeability, injection rate and formation pressure all affected the change in annular pressure between packers.

Published
2020
Full Text
View/download PDF

42. Correlation Between Apelin and Collateral Circulation in Patients with Middle Cerebral Artery Occlusion and Moyamoya Disease

Author

Wu,Hanlin, Xia,Chengyu, Li,Rui, Tao,Chunrong, Tang,Qiqiang, Hu,Wei, Wu,Hanlin, Xia,Chengyu, Li,Rui, Tao,Chunrong, Tang,Qiqiang, and Hu,Wei

Abstract

Hanlin Wu,1,* Chengyu Xia,2,* Rui Li,3 Chunrong Tao,3 Qiqiang Tang,3 Wei Hu1 1Stroke Center & Department of Neurology, The First Affiliated Hospital of USTC, Division of Life Sciences and Medicine, University of Science and Technology of China, Hefei, Anhui, People’s Republic of China; 2Department of Neurosurgery, The First Affiliated Hospital of USTC, Provincial Hospital Affiliated to Anhui Medical University, Hefei, Anhui, People’s Republic of China; 3Department of Neurology, The First Affiliated Hospital of USTC, Provincial Hospital Affiliated to Anhui Medical University, Hefei, Anhui, People’s Republic of China*These authors contributed equally to this workCorrespondence: Wei Hu; Qiqiang Tang Email andinghu@126.com; tqq1995@126.comBackground and Objectives: Moyamoya disease (MMD) is a unique cerebrovascular occlusive disease with abnormal vascular hyperplasia, which causes cerebrovascular accidents like intracranial arteriosclerosis. This study aimed to explore whether plasma apelin levels are related to good collateral circulation in ischemic diseases, which may be higher in patients with MMD than middle cerebral artery (MCA) occlusion or healthy controls, and may have a connection with the MMD grades.Methods: We recruited 68 MMD patients and 25 MCA occlusion patients diagnosed by angiography, including 29 patients without cerebrovascular problems as controls. We examined the plasma apelin, serum nitric oxide (NO), and vascular endothelial growth factor (VEGF) levels of all subjects by ELISA kit. We compared the relationship between apelin, NO, and VEGF in the blood of three groups, to explore the relationship. We also investigated whether the plasma apelin-13, apelin-17, and apelin-36 levels correlate with the MMD classification.Results: Univariate analyses indicated that the MMD group had the higher plasma apelin-13, apelin-17, apelin-36, and serum NO levels than the MCA occlusion and healthy control groups. Binary logist

Published
2022

49. Fatigue optimization of rotary control head rubber core based on steady sealing

Author

Huang Jian, Xueying Han, Junxiong Li, Lianglin Guo, Qian Liqin, Wang Zhiliang, Xia Chengyu, and Yijin Zeng

Subjects
Materials science, Yeoh, General Engineering, Drill pipe, Compression (physics), Finite element method, Core (optical fiber), Stress (mechanics), Natural rubber, visual_art, visual_art.visual_art_medium, Head (vessel), General Materials Science, Composite material
Abstract

In the process of underbalanced pressure drilling, the drill pipe body and its joint repeatedly pass though the sealing rubber core of the rotating blowout preventer (RBOP), making the rubber core susceptible to alternating cyclic load, and prone to fatigue failure. Under the premise of effective sealing realized by the rubber core, to increase its service life, first the Yeoh constitutive model and constitutive parameters describing the material of rubber core were determined by uniaxial tensile and compression tests, and the finite element governing equation of the nonlinear large deformation of the rubber core was obtained according to the principle of virtual work; then a three-dimensional finite element model of the rubber core was established by ABAQUS software to simulate the dynamic sealing process, contact pressure and Mises stress distribution and variation on the sealing surface were analyzed, and the accuracy of the finite element analysis results was verified by comparing the existing test results; subsequently, the influence of structural parameters such as inner angle, outer angle and inner diameter on the sealing surface stress was studied, and these parameters were optimized through Taguchi experiment. The optimal combination of structural parameters was obtained as inner angle 29°, outer angle 69° and inner diameter 84mm. The maximum Mises stress was reduced by 2.48MPa and the Mises stress amplitude by 42% after optimization. Finally, the suitable conditions of the optimized rubber core were selected based on the applicability analysis to conduct the field application and result shows that the rubber core has good sealing performance and significantly increased fatigue life after optimization.

Published
2022

Catalog

Books, media, physical & digital resources
See catalog results