Search

Your search keyword '"WANG, Guoqing"' showing total 39 results
Start Over Author "WANG, Guoqing" Publication Year Range This year
39 results on '"WANG, Guoqing"'

1. Directional Diffusion-Style Code Editing Pre-training

Author

Liang, Qingyuan, Sun, Zeyu, Zhu, Qihao, Hu, Junhao, Zhao, Yifan, Chen, Yizhou, Zhu, Mingxuan, Wang, Guoqing, and Zhang, Lu

Subjects
Computer Science - Software Engineering
Abstract

Code pre-trained models have shown promising effectiveness in various software engineering tasks. Among these tasks, many tasks are related to software evolution and/or code editing. However, existing code pre-trained models often overlook the real-world code editing data and the evolutionary nature of the editing process. In this paper, to simulate the step-by-step code editing process of human developers, we propose DivoT5, a pre-trained model based on directional diffusion at the data level. In DivoT5, we adopt two categories of pre-training tasks. The first category is mask and denoising tasks augmented with a diffusion direction representing code evolution. That is, we first apply a noising process to the code snippets before evolution, and then ask the pre-training process to restore the snippets with noise into the code snippets after evolution. The second category is tasks aiming to reinforce the evolutionary direction. That is, we first generate various intermediate versions for each pair of snippets before and after evolution, and then ask the pre-training process to transform the intermediate versions into the snippet after evolution for each pair. We evaluate DivoT5 for two code-editing scenarios and one non-editing scenario using five downstream tasks. Given each downstream task, we fine-tune the pre-trained DivoT5 to evaluate its effectiveness. Our experimental results show that DivoT5 achieves state-of-the-art (SOTA) performance on most tasks in comparison to models of the same scale (220M), large scale (770M) models in fine-tuning, and billion-scale (6.7B, 8B, ChatGPT) models in few-shot settings. For one code-editing task (i.e., automated code review), DivoT5 pre-trained on top of CodeT5-small (60M) can even outperform CodeT5-base (220M) and other pre-trained models with 220M parameters except for DivoT5 pre-trained on top of CodeT5-base (220M).

Published
2025

2. Disentangled Modeling of Preferences and Social Influence for Group Recommendation

Author

Ye, Guangze, Wu, Wen, Wang, Guoqing, Chen, Xi, Zheng, Hong, and He, Liang

Subjects
Computer Science - Information Retrieval
Abstract

The group recommendation (GR) aims to suggest items for a group of users in social networks. Existing work typically considers individual preferences as the sole factor in aggregating group preferences. Actually, social influence is also an important factor in modeling users' contributions to the final group decision. However, existing methods either neglect the social influence of individual members or bundle preferences and social influence together as a unified representation. As a result, these models emphasize the preferences of the majority within the group rather than the actual interaction items, which we refer to as the preference bias issue in GR. Moreover, the self-supervised learning (SSL) strategies they designed to address the issue of group data sparsity fail to account for users' contextual social weights when regulating group representations, leading to suboptimal results. To tackle these issues, we propose a novel model based on Disentangled Modeling of Preferences and Social Influence for Group Recommendation (DisRec). Concretely, we first design a user-level disentangling network to disentangle the preferences and social influence of group members with separate embedding propagation schemes based on (hyper)graph convolution networks. We then introduce a socialbased contrastive learning strategy, selectively excluding user nodes based on their social importance to enhance group representations and alleviate the group-level data sparsity issue. The experimental results demonstrate that our model significantly outperforms state-of-the-art methods on two realworld datasets., Comment: AAAI 2025 Oral

Published
2025

3. Automatically Learning a Precise Measurement for Fault Diagnosis Capability of Test Cases

Author

Zhao, Yifan, Sun, Zeyu, Wang, Guoqing, Liang, Qingyuan, Zhang, Yakun, Lou, Yiling, Hao, Dan, and Zhang, Lu

Subjects
Computer Science - Software Engineering
Abstract

Prevalent Fault Localization (FL) techniques rely on tests to localize buggy program elements. Tests could be treated as fuel to further boost FL by providing more debugging information. Therefore, it is highly valuable to measure the Fault Diagnosis Capability (FDC) of a test for diagnosing faults, so as to select or generate tests to better help FL. To this end, researchers have proposed many FDC metrics, which serve as the selection criterion in FL-oriented test selection or the fitness function in FL-oriented test generation. Existing FDC metrics can be classified into result-agnostic and result-aware metrics depending on whether they take test results (i.e., passing or failing) as input. Although result-aware metrics perform better in test selection, they have restricted applications due to the input of test results, e.g., they cannot be applied to guide test generation. Moreover, all the existing FDC metrics are designed based on some predefined heuristics and have achieved limited FL performance due to their inaccuracy. To address these issues, in this paper, we reconsider result-agnostic metrics, and propose a novel result-agnostic metric RLFDC which predicts FDC values of tests through reinforcement learning. In particular, we treat FL results as reward signals, and train an FDC prediction model with the direct FL feedback to automatically learn a more accurate measurement rather than design one based on predefined heuristics. Finally, we evaluate the proposed RLFDC on Defects4J by applying the studied metrics to test selection and generation. According to the experimental results, the proposed RLFDC outperforms all the result-agnostic metrics in both test selection and generation., Comment: This paper has been accepted by TOSEM

Published
2025

7. Design of a New Dual-Modal Ultrasonic Skull Cutting Device

Author

Dai, Zijun, Wang, Guoqing, Zhang, Qinghua, Wang, Shuai, Goos, Gerhard, Series Editor, Hartmanis, Juris, Founding Editor, Bertino, Elisa, Editorial Board Member, Gao, Wen, Editorial Board Member, Steffen, Bernhard, Editorial Board Member, Yung, Moti, Editorial Board Member, Lan, Xuguang, editor, Mei, Xuesong, editor, Jiang, Caigui, editor, Zhao, Fei, editor, and Tian, Zhiqiang, editor

Published
2025
Full Text
View/download PDF

8. Switching Sequence Control for Precise Reference Tracking of Power Converters

Author

Wang, Guoqing, Li, Guangqi, Dai, Zhiyong, Pan, Bowei, Angrisani, Leopoldo, Series Editor, Arteaga, Marco, Series Editor, Chakraborty, Samarjit, 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, Tan, Kay Chen, Series Editor, Yang, Qingxin, editor, Bie, Zhaohong, editor, and Yang, Xu, editor

Published
2025
Full Text
View/download PDF

9. ICPR 2024 Competition on Resource-Limited Infrared Small Target Detection Challenge: Methods and Results

Author

Li, Boyang, Ying, Xinyi, Li, Ruojing, Liu, Yongxian, Shi, Yangsi, Li, Miao, Zhang, Xin, Hu, Mingyuan, Wu, Chenyang, Zhang, Yukai, Wei, Hui, Tang, Dongli, Zhao, Jian, Jin, Lei, Xiao, Chao, Ling, Qiang, Lin, Zaiping, Sheng, Weidong, Li, Weixing, Peng, Chenxu, Ke, Wei, He, Yuhang, Yang, Huoren, Liu, Lingjie, Liu, Ziqi, Shi, Zelin, Liu, Yunpeng, Yu, Chuang, Zhao, Jinmiao, Yang, Shan, Xiang, Heng, Li, You, Wang, Guoqing, Li, Tianyu, Zhou, Minghang, Lan, Chenxi, Xi, Dongyu, Qiao, Chaofan, Gao, Yupeng, Liu, Yongxu, Feng, Wei, Ma, Zhihao, Chen, Deping, Song, Xiaopeng, Yang, Jiuping, Qiu, Zhaobing, Ni, Rixiang, Luo, Changhai, Lin, Ye, Zheng, Shuyuan, Huang, Baojin, Zhou, Xiaoqi, Guo, Qingshan, Zhang, Xuyang, Li, Xiuhong, Wu, Dangxuan, Ma, Jian, Zeng, Haodong, Wang, Luyao, Fu, Qiang, Dai, Yimian, Kou, Renke, Song, Jian, Feng, Changfeng, Xiong, Zihao, Xiao, Mengxuan, Huang, Hong, Liu, Yingxu, Zhao, Quanyi, Goos, Gerhard, Series Editor, Hartmanis, Juris, Founding Editor, Bertino, Elisa, Editorial Board Member, Gao, Wen, Editorial Board Member, Steffen, Bernhard, Editorial Board Member, Yung, Moti, Editorial Board Member, Antonacopoulos, Apostolos, editor, Chaudhuri, Subhasis, editor, Chellappa, Rama, editor, Liu, Cheng-Lin, editor, Bhattacharya, Saumik, editor, and Pal, Umapada, editor

Published
2025
Full Text
View/download PDF

10. OccGen: Generative Multi-modal 3D Occupancy Prediction for Autonomous Driving

Author

Wang, Guoqing, Wang, Zhongdao, Tang, Pin, Zheng, Jilai, Ren, Xiangxuan, Feng, Bailan, Ma, Chao, Goos, Gerhard, Series Editor, Hartmanis, Juris, Founding Editor, Bertino, Elisa, Editorial Board Member, Gao, Wen, Editorial Board Member, Steffen, Bernhard, Editorial Board Member, Yung, Moti, Editorial Board Member, Leonardis, Aleš, editor, Ricci, Elisa, editor, Roth, Stefan, editor, Russakovsky, Olga, editor, Sattler, Torsten, editor, and Varol, Gül, editor

Published
2025
Full Text
View/download PDF

11. Diffusion Models as Optimizers for Efficient Planning in Offline RL

Author

Huang, Renming, Pei, Yunqiang, Wang, Guoqing, Zhang, Yangming, Yang, Yang, Wang, Peng, Shen, Hengtao, Goos, Gerhard, Series Editor, Hartmanis, Juris, Founding Editor, Bertino, Elisa, Editorial Board Member, Gao, Wen, Editorial Board Member, Steffen, Bernhard, Editorial Board Member, Yung, Moti, Editorial Board Member, Leonardis, Aleš, editor, Ricci, Elisa, editor, Roth, Stefan, editor, Russakovsky, Olga, editor, Sattler, Torsten, editor, and Varol, Gül, editor

Published
2025
Full Text
View/download PDF

12. Region-Aware Distribution Contrast: A Novel Approach to Multi-task Partially Supervised Learning

Author

Li, Meixuan, Li, Tianyu, Wang, Guoqing, Wang, Peng, Yang, Yang, Zou, Jie, Goos, Gerhard, Series Editor, Hartmanis, Juris, Founding Editor, Bertino, Elisa, Editorial Board Member, Gao, Wen, Editorial Board Member, Steffen, Bernhard, Editorial Board Member, Yung, Moti, Editorial Board Member, Leonardis, Aleš, editor, Ricci, Elisa, editor, Roth, Stefan, editor, Russakovsky, Olga, editor, Sattler, Torsten, editor, and Varol, Gül, editor

Published
2025
Full Text
View/download PDF

31. Properties of tartaric acid modified steel slag as supplementing cementitious materials.

Author

Tang, Panpan, Xiong, Yangkai, Huang, Lei, Fang, Zhiqiang, Jiang, Hao, and Wang, Guoqing

Subjects
TARTARIC acid, CALCIUM silicate hydrate, LIME (Minerals), CALCIUM hydroxide, FLEXURAL strength, MORTAR, CALCIUM silicates
Abstract

By using tartaric acid (TA) as a wet‐method modifier to modify steel slag powder (SS), it is helpful to reduce the content of free calcium oxide (f‐CaO) in SS and further improve the mechanical properties and soundness of SS as supplementing cementitious materials (SCM) in the mortar. The results indicate that with the addition of 2 wt% of TA, the number of large particles bigger than 45 µm in the SS decreased, resulting in predominantly smaller particles smaller than 45 µm. While the specific surface area increased from 417 to 704 m2/kg, the water consumption at standard consistency was significantly reduced. The soundness of the paste SS as SCM had improved including the f‐CaO content decreased from 4.81% to 0.95%, and the Le Chatelier expansion reduced from 4.5 to 1.5 mm. The mechanical properties were significantly enhanced, with the flexural strength increasing from 5.6 to 7.8 MPa, and the compressive strength rising from 38.8 to 52.7 MPa. After 28 days of curing in water, the hydration products of the mortar are hydrated calcium silicate (C‐S‐H), calcium hydroxide (CH), and calcium carbonate (CaCO3). [ABSTRACT FROM AUTHOR]

Published
2025
Full Text
View/download PDF

32. Elements gradient doping in Mn-based Li-rich layered oxides for long-life lithium-ion batteries.

Author

Wang, Yinzhong, Liu, Shiqi, Guo, Xianwei, Wang, Boya, Zhang, Qinghua, Li, Yuqiang, Wang, Yulong, Wang, Guoqing, Gu, Lin, and Yu, Haijun

Subjects
ENERGY density, ENERGY storage, INTERFACE stability, STORAGE batteries, INTERFACE structures
Abstract

• The cobalt-free Mn-LLOs cathode materials with a small amount of Li 2 MnO 3 crystal domain in the bulk structure and gradient doping of Al and Ti elements have been developed. • The batteries with Mn-LLOs cathode deliver the energy density over 600 Wh kg−1, excellent cycle retention of 99.7 % after 800 cycles with low voltage decay of 0.03 mV cycle−1. • The Mn-LLOs cathode materials with bulk structure design and interface regulation can improve the long-time cycling stability of low-cost Mn-based energy storage batteries. The cobalt-free Mn-based Li-rich layered oxide material has the advantages of low cost, high energy density, and good performance at low temperatures, and is the promising choice for energy storage batteries. However, the long-cycling stability of batteries needs to be improved. Herein, the Mn-based Li-rich cathode materials with small amounts of Li 2 MnO 3 crystal domains and gradient doping of Al and Ti elements from the surface to the bulk have been developed to improve the structure and interface stability. Then the batteries with a high energy density of 600 Wh kg−1, excellent capacity retention of 99.7 % with low voltage decay of 0.03 mV cycle−1 after 800 cycles, and good rates performances can be achieved. Therefore, the structure and cycling stability of low voltage Mn-based Li-rich cathode materials can be significantly improved by the bulk structure design and interface regulation, and this work has paved the way for developing low-cost and high-energy Mn-based energy storage batteries with long lifetime. [Display omitted] [ABSTRACT FROM AUTHOR]

Published
2025
Full Text
View/download PDF

35. Trends in industrial ethylene production: Innovation in process and catalyst design

Author

Jiang, Chaoran, He, Feng, Zhang, Lijun, Wang, Guoqing, and Liu, Lichen

Abstract

In this Activity article, Prof. Lichen Liu (associate professor at Tsinghua University) and Prof. Guoqing Wang (chief research fellow of the SINOPEC group at the SINOPEC [Beijing] Research Institute of Chemical Industry) exchange views from industrial and academic perspectives on the trends in ethylene production in the chemical industry and the current challenges in developing advanced oxide- and zeolite-based catalysts for ethylene production through catalytic cracking. Furthermore, they give perspectives on the alternative processes for ethylene production and the promising directions in catalyst design and process engineering.

Published
2025
Full Text
View/download PDF

36. Opportunities and challenges in the ethylene value chain

Author

He, Feng, Jiang, Chaoran, Zhang, Lijun, Wang, Guoqing, and Liu, Lichen

Abstract

In this Activity article, Prof. Guoqing Wang (chief research fellow of the SINOPEC group at the SINOPEC [Beijing] Research Institute of Chemical Industry) and Prof. Lichen Liu (associate professor at Tsinghua University) exchange views from industrial and academic perspectives on the current status of the ethylene value chain and discuss the emerging trends in the downstream markets. Furthermore, they give perspectives on the opportunities and challenges in designing efficient catalysts for the downstream processes for converting ethylene into value-added chemicals and materials.

Published
2025
Full Text
View/download PDF

37. Phosphorylation of PA at serine 225 enhances viral fitness of the highly pathogenic H5N1 avian influenza virus in mice.

Author

Zhang M, Zeng Z, Chen X, Wang G, Cai X, Hu Z, Gu M, Hu S, Liu X, Wang X, Peng D, Hu J, and Liu X

Abstract

Currently, there is increasing spillover of highly pathogenic H5N1 avian influenza virus (AIV) to mammals, raising a concern of pandemic threat about this virus. Although the function of PA protein of the influenza virus is well understood, the understanding of how phosphorylation regulates this protein and influenza viral life cycle is still limited. We previously identified PA S225 as the phosphorylation site in the highly pathogenic H5N1 AIV. In this study, we investigated the role of phosphorylation in regulating PA function and viral fitness through dephosphorylation (PA S225A) or continuous phosphorylation (PA S225E)-mimetic mutation of PA S225. Structure analysis revealed that PA S225A or PA S225E mutation had no obvious effect on the structure of PA protein. Replication assay in vitro showed that PA S225A phosphorylation-ablative mutation significantly inhibited virus replication both in mammalian and avian-derived cells, while PA S225E enhanced viral replication in these cells. Correspondingly, PA S225A dephosphorylation significantly attenuated viral replication and virulence in mice, while PA S225E enhanced these aspects in mice. Mechanistically, PA S225A mutation significantly decreased viral polymerase activity, disabled viral ribonucleoprotein complex (vRNP) assembly and attenuated PA nuclear accumulation. Altogether, our study directly suggested that phosphorylation of PA protein at site S225 enhances viral fitness of the highly pathogenic H5N1 virus in mammals by assuring effective vRNP activity, providing a framework for further study of phosphorylation events in influenza virus life cycle., Competing Interests: Declaration of Competing Interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper., (Copyright © 2025 Elsevier B.V. All rights reserved.)

Published
2025
Full Text
View/download PDF

38. FOXM1 promotes malignant biological behavior and metabolic reprogramming by targeting SPINK1 in hepatocellular carcinoma and affecting the p53 pathway.

Author

Ding X, Shi J, Lei Z, Wang G, Fu C, Su X, and Zhu G

Abstract

This study investigates the role of SPINK1 in liver cancer and its regulatory relationship with FOXM1. Using differential gene analysis in the GEO database, SPINK1 was identified as overexpressed in liver cancer tissues and associated with poor prognosis, confirmed via PCR. Functional assays demonstrated that SPINK1 knockdown reduced proliferation, migration, and invasion in liver cancer cells, while promoting apoptosis. In vivo experiments revealed that SPINK1 knockdown inhibited tumor growth, decreased Ki-67 and N-cadherin levels, increased E-cadherin levels, and suppressed lung metastasis. Analysis of upstream factors indicated that FOXM1 binds to the SPINK1 promoter, as validated by dual-luciferase and ChIP assays, thereby promoting SPINK1 transcription. TCGA database analysis and clinical tissue validation showed that FOXM1 expression correlates with poor prognosis in liver cancer. Functional studies demonstrated that FOXM1 knockdown suppressed liver cancer progression, while SPINK1 overexpression reversed these effects. KEGG enrichment analysis identified the p53 pathway as a key downstream target of SPINK1, and Western blotting confirmed its role in modulating p53 pathway activity. These findings reveal a critical FOXM1-SPINK1 axis in liver cancer progression. FOXM1 directly promotes SPINK1 transcription, enhancing tumor cell proliferation and metastasis while regulating the p53 pathway. Targeting this axis could provide a potential therapeutic approach for liver cancer., Competing Interests: Declaration of competing interest None declared., (Copyright © 2025. Published by Elsevier B.V.)

Published
2025
Full Text
View/download PDF

39. Identification and validation of Atp5f1c in CD4 + T cell as a hub protein in Parkinson's disease.

Author

Wang G, Wang Y, Tang X, Li D, Zhao Y, and Zhang F

Abstract

Parkinson's disease (PD) is an age-related and progressive neurodegenerative disease. Growing evidences indicate that CD4 + T cell dysfunction plays an essential role in the progress of PD. Here, in LPS-induced PD mice, we isolated midbrain CD4 + T cell and peripheral CD4 + T cell to perform proteomics, and then screened a total of 167 co-expression proteins via integrated bioinformatics analysis. In addition, the subcellular localization, GO analysis, KEGG pathways and protein-protein interaction of 167 co-expression proteins were assessed. Furthermore, GeneMANIA searched the hub proteins and their co-expression genes and found 13 overlapping hub proteins, including Ndufa3, Cox5b, Mrpl21, Ndufab1, Idh3g, Ndufb7, Cyc1, Cisd1, Atp5f1c, Sdhc, Ndufb9, Mtnd1 and Mrpl17. Next, GO analysis and KEGG analysis of the 13 overlapping hub proteins were also exhibited. Further analysis identified that 4 hub proteins (Idh3g, Cisd1, Atp5f1c and Mtnd1) were downregulated both in midbrain and peripheral CD4 + T cell from proteomics. Identification and rescue experiment analysis showed that only Atp5f1c was decreased in LPS- and 6-OHDA-induced PD mice and dopamine (DA) neuronal loss and ATP production decrease were disappeared after Atp5f1c over-expression/Atp5f1c reinfusion both in vivo and in vitro. In conclusion, Atp5f1c was verified as a potential CD4 + T cell-related hub protein for PD., Competing Interests: Declaration of competing interest The authors have declared no conflict of interest., (Copyright © 2025 Elsevier B.V. All rights reserved.)

Published
2025
Full Text
View/download PDF

Catalog

Books, media, physical & digital resources
See catalog results