Search

Your search keyword '"Wang, Leilei"' showing total 1,331 results
Start Over Author "Wang, Leilei" Search Limiters Peer Reviewed Publication Year Range Last 10 years
1,331 results on '"Wang, Leilei"'

1. Solution for CVPR 2024 UG2+ Challenge Track on All Weather Semantic Segmentation

Author

Yu, Jun, Zhang, Yunxiang, Sun, Fengzhao, Wang, Leilei, and Lu, Renjie

Subjects
Computer Science - Computer Vision and Pattern Recognition, Computer Science - Artificial Intelligence
Abstract

In this report, we present our solution for the semantic segmentation in adverse weather, in UG2+ Challenge at CVPR 2024. To achieve robust and accurate segmentation results across various weather conditions, we initialize the InternImage-H backbone with pre-trained weights from the large-scale joint dataset and enhance it with the state-of-the-art Upernet segmentation method. Specifically, we utilize offline and online data augmentation approaches to extend the train set, which helps us to further improve the performance of the segmenter. As a result, our proposed solution demonstrates advanced performance on the test set and achieves 3rd position in this challenge., Comment: Solution for CVPR 2024 UG2+ Challenge Track on All Weather Semantic Segmentation

Published
2024

10. RRWKV: Capturing Long-range Dependencies in RWKV

Author

Wang, Leilei

Subjects
Computer Science - Computation and Language, Computer Science - Artificial Intelligence
Abstract

Owing to the impressive dot-product attention, the Transformers have been the dominant architectures in various natural language processing (NLP) tasks. Recently, the Receptance Weighted Key Value (RWKV) architecture follows a non-transformer architecture to eliminate the drawbacks of dot-product attention, where memory and computational complexity exhibits quadratic scaling with sequence length. Although RWKV has exploited a linearly tensor-product attention mechanism and achieved parallelized computations by deploying the time-sequential mode, it fails to capture long-range dependencies because of its limitation on looking back at previous information, compared with full information obtained by direct interactions in the standard transformer. Therefore, the paper devises the Retrospected Receptance Weighted Key Value (RRWKV) architecture via incorporating the retrospecting ability into the RWKV to effectively absorb information, which maintains memory and computational efficiency as well.

Published
2023

12. Research on Full Digital Control High-voltage Direct Current LLC Power Supply for Rail Transit Based on Domestic Embedded Chips

Author

XUE Hongquan, CHANG Jianglong, LI Jiacheng, and WANG Leilei

Subjects
rail transit, high-voltage dc power supply, llc resonant converter, primary harmonic approximation method, Transportation engineering, TA1001-1280
Abstract

Objective High-efficiency LLC (logic link control of) topology high-voltage direct current power supplies are widely used in the field of rail transit. However, due to the variable frequency control characteristics of LLC circuits, achieving full digitalization is challenging, and research on domestication lags behind. Thus it is necessary to carry out relevant research. Method Based on the HPM6300 series of full digital control LLC high-voltage power supply circuits from Shanghai Xianji Semiconductor Technology Co., Ltd. (hereinafter referred to as 'Xianji'), an analysis is conducted using the primary harmonic approximation method to establish a circuit model. The control flow of the circuit system program is analyzed. Experimental verification is performed by constructing a prototype with a power of 1.5 kW. Result & Conclusion The experimental results confirm that using the domestic Xianji embedded control chip HPM6300 MCU to build an LLC test prototype, and through PFM+PWM hybrid control, wide-range input can be achieved, and voltage output can be quickly stabilized. Different control methods can be adopted in the future, to achieve wide-range input with structural changes.

Published
2024
Full Text
View/download PDF

14. Automatic Voice Quality Evaluation Method of lVR Service in Call Center Based on Stacked Auto Encoder

Author

Wang, Li, Wang, Zongwei, Zhao, Guoyi, Su, Yuan, Zhao, Jinli, Wang, Leilei, Angrisani, Leopoldo, Series Editor, Arteaga, Marco, Series Editor, Chakraborty, Samarjit, Series Editor, Chen, Jiming, Series Editor, Chen, Shanben, Series Editor, Chen, Tan Kay, Series Editor, Dillmann, Rüdiger, Series Editor, Duan, Haibin, Series Editor, Ferrari, Gianluigi, Series Editor, Ferre, Manuel, Series Editor, Jabbari, Faryar, Series Editor, Jia, Limin, Series Editor, Kacprzyk, Janusz, Series Editor, Khamis, Alaa, Series Editor, Kroeger, Torsten, Series Editor, Li, Yong, Series Editor, Liang, Qilian, Series Editor, Martín, Ferran, Series Editor, Ming, Tan Cher, Series Editor, Minker, Wolfgang, Series Editor, Misra, Pradeep, Series Editor, Mukhopadhyay, Subhas, Series Editor, Ning, Cun-Zheng, Series Editor, Nishida, Toyoaki, Series Editor, Oneto, Luca, Series Editor, Panigrahi, Bijaya Ketan, Series Editor, Pascucci, Federica, Series Editor, Qin, Yong, Series Editor, Seng, Gan Woon, Series Editor, Speidel, Joachim, Series Editor, Veiga, Germano, Series Editor, Wu, Haitao, Series Editor, Zamboni, Walter, Series Editor, Zhang, Junjie James, Series Editor, Tan, Kay Chen, Series Editor, Hung, Jason C., editor, Yen, Neil, editor, and Chang, Jia-Wei, editor

Published
2024
Full Text
View/download PDF

15. Innovative ASR Technology for Profile Contour and Flatness Control of Electrical Steel in Multiple-Width Schedule Free Rolling

Author

Cao, Jianguo, Song, Chunning, Wang, Leilei, Zhao, Qiufang, Xiao, Jing, Sun, Lei, Chaari, Fakher, Series Editor, Gherardini, Francesco, Series Editor, Ivanov, Vitalii, Series Editor, Haddar, Mohamed, Series Editor, Cavas-Martínez, Francisco, Editorial Board Member, di Mare, Francesca, Editorial Board Member, Kwon, Young W., Editorial Board Member, Trojanowska, Justyna, Editorial Board Member, Xu, Jinyang, Editorial Board Member, Mocellin, Katia, editor, Bouchard, Pierre-Olivier, editor, Bigot, Régis, editor, and Balan, Tudor, editor

Published
2024
Full Text
View/download PDF

18. ESIE-BERT: Enriching Sub-words Information Explicitly with BERT for Joint Intent Classification and SlotFilling

Author

Guo, Yu, Xie, Zhilong, Chen, Xingyan, Chen, Huangen, Wang, Leilei, Du, Huaming, Wei, Shaopeng, Zhao, Yu, Li, Qing, and Wu, Gang

Subjects
Computer Science - Computation and Language, Computer Science - Artificial Intelligence
Abstract

Natural language understanding (NLU) has two core tasks: intent classification and slot filling. The success of pre-training language models resulted in a significant breakthrough in the two tasks. One of the promising solutions called BERT can jointly optimize the two tasks. We note that BERT-based models convert each complex token into multiple sub-tokens by wordpiece algorithm, which generates a mismatch between the lengths of the tokens and the labels. This leads to BERT-based models do not do well in label prediction which limits model performance improvement. Many existing models can be compatible with this issue but some hidden semantic information is discarded in the fine-tuning process. We address the problem by introducing a novel joint method on top of BERT which explicitly models the multiple sub-tokens features after wordpiece tokenization, thereby contributing to the two tasks. Our method can well extract the contextual features from complex tokens by the proposed sub-words attention adapter (SAA), which preserves overall utterance information. Additionally, we propose an intent attention adapter (IAA) to obtain the full sentence features to aid users to predict intent. Experimental results confirm that our proposed model is significantly improved on two public benchmark datasets. In particular, the slot filling F1 score is improved from 96.1 to 98.2 (2.1% absolute) on the Airline Travel Information Systems (ATIS) dataset.

Published
2022

19. Normalized evaluation for wire arc additive manufacturing of 2319 aluminum alloy

Author

LYU Feiyue, WANG Leilei, DOU Zhiwei, LIU Shengxin, DU Mingzhen, GAO Chuanyun, and ZHAN Xiaohong

Subjects
wire arc additive manufacturing, 2319 aluminium alloy, corresponding rule, normalized evaluation, Materials of engineering and construction. Mechanics of materials, TA401-492
Abstract

To analyze quantitatively various indexes of wire arc additive manufacturing components of 2319 aluminum alloy, the expressions of curved surface between different process parameters, porosity and tensile strength values were fitted, and the "process-structure-property" corresponding rule was established. In addition, a normalized fuzzy evaluation model was developed for process parameters, organizational defects, and mechanical properties using the generalized fuzzy synthesis operation rule. The purpose of this model was to obtain the optimal process parameters. The results show that the porosity generally increases with the wire feeding speed rising.Then the porosity generally decreases with a decrease in scanning speed. When the scanning speed is 0.035 m/s, the correlation between porosity and tensile strength is the lowest, with a correlation coefficient (coefficient of determination, COD) of only 0.6. The comprehensive evaluation score of the expert is the highest when the wire feeding speed is 5.0 m/min and the scanning speed is 0.025 m/s, indicating that this combination of process parameters is optimal.

Published
2024
Full Text
View/download PDF

30. Male-specific, replicable and functional roles of genetic variants and cerebral gray matter volumes in ADHD: a gene-wide association study across KTN1 and a region-wide functional validation across brain

Author

Luo, Xingguang, Lin, Xiandong, Ide, Jaime S., Luo, Xinqun, Zhang, Yong, Xu, Jianying, Wang, Leilei, Chen, Yu, Cheng, Wenhong, Zheng, Jianming, Wang, Zhiren, Yu, Ting, Taximaimaiti, Reyisha, Jing, Xiaozhong, Wang, Xiaoping, Cao, Yuping, Tan, Yunlong, and Li, Chiang-Shan R.

Published
2023
Full Text
View/download PDF

46. Effects of Swirl Distortion on Flow Field and Blade Load of a Centrifugal Compressor

Author

Li Yanzhao, Hou Hongjuan, and Wang Leilei

Subjects
swirl distortion, centrifugal compressor, leakage flow, blade load, 97p10, Mathematics, QA1-939
Abstract

On the turbocharger, two forms of intake pipe, bending and torsion, are usually used, which leads to a vortex at the inlet of the blade. It has been found that double vortices and massive vortices play a significant role in the working characteristics of the compressor. This project intends to conduct a systematic theoretical and experimental study on the effects of vortex deformation under two different working conditions of bending and torsion on the internal flow field, tip leakage flow field, and impeller load of the impeller. It has been found that the presence of a vortex near the blade entrance changes the static pressure and inflow angle near the blade entrance. The reverse transfer of vortex flow and vortex flow caused by vortex has three critical effects on the working characteristics of the compressor: 1. The change of circumferential static pressure and the pressure gradient of the flow in the impeller. 2. The location and density of the leakage flow on the top surface are significantly changed. 3. Adjust the impeller load’s change and fluctuation when it rotates.

Published
2024
Full Text
View/download PDF

47. Unsteady Flow Characteristics in a Centrifugal Compressor

Author

Li Yanzhao, Hou Hongjuan, and Wang Leilei

Subjects
centrifugal compressor, unsteady flow, jet-wake structure, flow parameters, blade load, 97p10, Mathematics, QA1-939
Abstract

The geometry of the vortex shell significantly affects the non-axisymmetric flow in centrifugal compressors. This study explores the static pressure characteristics of the vortex shell through experimental data and numerical analysis using finite element software. Findings indicate that the vortex tongue region critically influences the cascade structure. At high flow rates (0.40 kg/s), flow field parameters at the blade and vortex shell inlets alter markedly near the tongue side. Conversely, at lower velocities (0.26 kg/s), flow distribution at the blade inlet becomes more uneven, and changes in the inflow angle are more pronounced. Additionally, blade pulsation intensifies at higher velocities, and the jet wake vortex exhibits structural variations at different velocities. High-pressure fluctuations are notably transmitted upwards through the vortex shell, with periodic fluctuations affecting the impeller load. These changes significantly impact the flow field under various operating conditions.

Published
2024
Full Text
View/download PDF

Catalog

Books, media, physical & digital resources
See catalog results