You Zhang
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
You Zhang has a total of 25 co-authors. Here are the top co-authors by number of shared publications:
- Jin Wang
School of Information Science and Engineering, Yunnan University, Kunming, China
7 shared publications - Xuejie Zhang
School of Information Science and Engineering, Yunnan University, Kunming, China
7 shared publications Jin Wang
Yunnan University
5 shared publicationsXuejie Zhang
Yunnan University
4 shared publicationsDan Xu
Yunnan University
3 shared publications- Dan Xu
Yunnan University
3 shared publications
Latest Publications
- 17 November 2025
- Published by Springer Nature
- p. 488-499
Qwen-Gender: A Chain-of-Thought Based Multi-task Gender Bias Mitigation System
- 17 November 2025
- Published by Springer Nature
- p. 488-499
Qwen-Gender: A Chain-of-Thought Based Multi-task Gender Bias Mitigation System
- 18 December 2024
- Published by Elsevier in Information Sciences
- Vol. 698
Heterogeneous federated distillation with mutual information maximization for medical relation extraction
- 18 December 2024
- Published by Elsevier in Information Sciences
- Vol. 698
Heterogeneous federated distillation with mutual information maximization for medical relation extraction
- 26 March 2024
- Published by Elsevier in Neural Networks
- Vol. 175, 106272
Layerwised multimodal knowledge distillation for vision-language pretrained model
- 26 March 2024
- Published by Elsevier in Neural Networks
- Vol. 175, 106272
Layerwised multimodal knowledge distillation for vision-language pretrained model
- 18 March 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Emerging Topics in Computational Intelligence
- Vol. 8 (3) , 2581-2591
Attribute-Based Injection Transformer for Personalized Sentiment Analysis
Personal attributes have been proven to be useful for sentiment analysis. However, previous models of learning attribute-specific language representations are suboptimal because only context- or content-wise injection is adopted. This study proposes a transformer structure with a combination of both context- and content-wise injections based on a well pretrained transformer encoder. For context-wise injection, self-interactive attention is implemented by incorporating personal attributes into a multi-head attention. For the content-wise perspective, an attribute-based layer normalization is used to align text representation with personal attributes. In particular, the proposed transformer layer can be a universal layer compatible with the original Google Transformer layer. Instead of training from scratch, the proposed Transformer layer can be initialized from a well pre-trained checkpoint for downstream tasks. Extensive experiments were conducted on three benchmarks of document-level sentiment analysis, including IMDB, Yelp-2013 and Yelp-2014. The results show that the proposed method outperforms the previous methods for personalized sentiment analysis, demonstrating that the combination of both context- and content-wise injections can facilitate model learning for attribute-specific language representations.
- 18 March 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Emerging Topics in Computational Intelligence
- Vol. 8 (3) , 2581-2591
Attribute-Based Injection Transformer for Personalized Sentiment Analysis
Personal attributes have been proven to be useful for sentiment analysis. However, previous models of learning attribute-specific language representations are suboptimal because only context- or content-wise injection is adopted. This study proposes a transformer structure with a combination of both context- and content-wise injections based on a well pretrained transformer encoder. For context-wise injection, self-interactive attention is implemented by incorporating personal attributes into a multi-head attention. For the content-wise perspective, an attribute-based layer normalization is used to align text representation with personal attributes. In particular, the proposed transformer layer can be a universal layer compatible with the original Google Transformer layer. Instead of training from scratch, the proposed Transformer layer can be initialized from a well pre-trained checkpoint for downstream tasks. Extensive experiments were conducted on three benchmarks of document-level sentiment analysis, including IMDB, Yelp-2013 and Yelp-2014. The results show that the proposed method outperforms the previous methods for personalized sentiment analysis, demonstrating that the combination of both context- and content-wise injections can facilitate model learning for attribute-specific language representations.
- Open Access
- 1 January 2024
- Published by Association for Computational Linguistics (ACL)
- p. 96-101
YNU-HPCC at SIGHAN-2024 dimABSA Task: Using PLMs with a Joint Learning Strategy for Dimensional Intensity Prediction
Wangzehui@stu.ynu.edu.cn Wangzehui@stu.ynu.edu.cn, You Zhang, Jin Wang, Dan Xu, Xuejie Zhang. Proceedings of the 10th SIGHAN Workshop on Chinese Language Processing (SIGHAN-10). 2024.
- 1 January 2024
- Published by Association for Computational Linguistics (ACL)
- p. 96-101
Open AccessYNU-HPCC at SIGHAN-2024 dimABSA Task: Using PLMs with a Joint Learning Strategy for Dimensional Intensity Prediction
Wangzehui@stu.ynu.edu.cn Wangzehui@stu.ynu.edu.cn, You Zhang, Jin Wang, Dan Xu, Xuejie Zhang. Proceedings of the 10th SIGHAN Workshop on Chinese Language Processing (SIGHAN-10). 2024.
- 17 November 2023
- Published by Springer Nature in Applied Intelligence
- Vol. 53 (24) , 30416-30430
Graphs get personal: learning representation with contextual pretraining for collaborative filtering
- 17 November 2023
- Published by Springer Nature in Applied Intelligence
- Vol. 53 (24) , 30416-30430
Graphs get personal: learning representation with contextual pretraining for collaborative filtering
- Open Access
- 7 July 2023
- Published by Association for Computing Machinery (ACM)
- p. 180-183
Contrastive Learning with Soft Negative Sampling for Visual Word Sense Disambiguation
The task of visual word sense disambiguation (VWSD) is to find the images that accurately match target words or phrases from a given candidate set of images. It is a challenging problem, requiring VWSD models to understand the related information between images and texts and distinguish the ambiguous information. To address this problem, we propose a method of contrastive learning with soft negative sampling (CLSNS), which facilitates the optimization of VWSD models via comparisons of positive and soft negative samples. When calculating the text and image comparison loss, the ambiguous target word is used as the soft negative sample, the target phrase with known semantics is used as the positive sample, and other irrelevant text image pairs are used as negative samples. bidirectional margin loss (BML) is utilized to distinguish between negative samples and soft negative samples, and Info Noise-contrastive estimation (InfoNCE) loss to distinguish between positive and negative samples. To make the model could be trained more effectively, we expand the target word and target phrase using prompt technology. We can broaden the diversity of the training data by inserting various prompts which will enhance the model's functionality and generalizability. These augmentations can improve the model's comprehension of and ability to provide domain-specific content, resulting in more precise and focused responses. Moreover, an extended Wikipedia-based text model is introduced to enrich the input text information. Extensive experiments conducted on a VWSD dataset shared from SemEval 2023 demonstrate the effect of the proposed method.
- 7 July 2023
- Published by Association for Computing Machinery (ACM)
- p. 180-183
Open AccessContrastive Learning with Soft Negative Sampling for Visual Word Sense Disambiguation
The task of visual word sense disambiguation (VWSD) is to find the images that accurately match target words or phrases from a given candidate set of images. It is a challenging problem, requiring VWSD models to understand the related information between images and texts and distinguish the ambiguous information. To address this problem, we propose a method of contrastive learning with soft negative sampling (CLSNS), which facilitates the optimization of VWSD models via comparisons of positive and soft negative samples. When calculating the text and image comparison loss, the ambiguous target word is used as the soft negative sample, the target phrase with known semantics is used as the positive sample, and other irrelevant text image pairs are used as negative samples. bidirectional margin loss (BML) is utilized to distinguish between negative samples and soft negative samples, and Info Noise-contrastive estimation (InfoNCE) loss to distinguish between positive and negative samples. To make the model could be trained more effectively, we expand the target word and target phrase using prompt technology. We can broaden the diversity of the training data by inserting various prompts which will enhance the model's functionality and generalizability. These augmentations can improve the model's comprehension of and ability to provide domain-specific content, resulting in more precise and focused responses. Moreover, an extended Wikipedia-based text model is introduced to enrich the input text information. Extensive experiments conducted on a VWSD dataset shared from SemEval 2023 demonstrate the effect of the proposed method.
- Open Access
- 1 January 2023
- Published by Clausius Scientific Press, Inc. in Advances in Vocational and Technical Education
- Vol. 5 (7) , 48-55
Current Situation and Path Optimization of College Students' Innovation and Entrepreneurship under the Background of "Internet Plus"
With the integration of the awareness of "Internet plus" ( "Internet+" ) into people's lives, the era of "Internet plus" has come. From the perspective of "Internet plus", starting with the opportunities and challenges of college students' innovation and entrepreneurship, this paper proposes an optimization path to solve practical problems. This paper probes into the current situation and path optimization of college students' innovation and entrepreneurship under the background of "Internet plus", expounds the research on the current situation of college students' entrepreneurship mode under the background of "Internet plus", finds the problems existing in college students' innovation and entrepreneurship under the background of "Internet plus", and puts forward the path optimization of college students' innovation and entrepreneurship under the background of "Internet plus". After optimizing the innovation and entrepreneurship path for college students, their cognitive ability, innovation ability, and innovation literacy continue to improve over time. Throughout the entire process, their cognitive ability has increased by 0.17; their innovation ability has increased by 0.20; their innovation literacy has increased by 0.17. After optimizing the innovation and entrepreneurship path, it can help students combine theory with practice, improve their practical ability and comprehensive quality, and also cultivate their thinking ability, laying a better foundation for college students' entrepreneurship.
- 1 January 2023
- Published by Clausius Scientific Press, Inc. in Advances in Vocational and Technical Education
- Vol. 5 (7) , 48-55
Open AccessCurrent Situation and Path Optimization of College Students' Innovation and Entrepreneurship under the Background of "Internet Plus"
With the integration of the awareness of "Internet plus" ( "Internet+" ) into people's lives, the era of "Internet plus" has come. From the perspective of "Internet plus", starting with the opportunities and challenges of college students' innovation and entrepreneurship, this paper proposes an optimization path to solve practical problems. This paper probes into the current situation and path optimization of college students' innovation and entrepreneurship under the background of "Internet plus", expounds the research on the current situation of college students' entrepreneurship mode under the background of "Internet plus", finds the problems existing in college students' innovation and entrepreneurship under the background of "Internet plus", and puts forward the path optimization of college students' innovation and entrepreneurship under the background of "Internet plus". After optimizing the innovation and entrepreneurship path for college students, their cognitive ability, innovation ability, and innovation literacy continue to improve over time. Throughout the entire process, their cognitive ability has increased by 0.17; their innovation ability has increased by 0.20; their innovation literacy has increased by 0.17. After optimizing the innovation and entrepreneurship path, it can help students combine theory with practice, improve their practical ability and comprehensive quality, and also cultivate their thinking ability, laying a better foundation for college students' entrepreneurship.
- Open Access
- 1 January 2023
- Published by Clausius Scientific Press, Inc. in Journal of Human Resource Development
- Vol. 5 (2) , 18-24
Branch Financial Technology Personnel Training System
In response to the problems of inconsistent training content and personnel training needs, as well as unreasonable training evaluation mechanisms in the training system of branch financial technology personnel, it is urgent to effectively optimize the personnel training system. This article analyzed the problems existing in the existing personnel training system, proposed corresponding optimization strategies for their specific problems, and further conducted a questionnaire survey on 8 financial technology personnel in a certain branch to test their satisfaction and practicality ratings before and after the optimization of the training system. Before optimizing the training system, the average satisfaction score of each technical personnel was 6.70 points, and the average practicality score was 5.59 points; After optimizing the training system, the average satisfaction score of each technical personnel was 8.56 points, and the average practicality score was 7.99 points. From the above data, it can be seen that after optimizing the personnel training system, technology personnel have higher satisfaction and practicality ratings for the training system, indicating that optimizing the personnel training system has certain practical significance.
- 1 January 2023
- Published by Clausius Scientific Press, Inc. in Journal of Human Resource Development
- Vol. 5 (2) , 18-24
Open AccessBranch Financial Technology Personnel Training System
In response to the problems of inconsistent training content and personnel training needs, as well as unreasonable training evaluation mechanisms in the training system of branch financial technology personnel, it is urgent to effectively optimize the personnel training system. This article analyzed the problems existing in the existing personnel training system, proposed corresponding optimization strategies for their specific problems, and further conducted a questionnaire survey on 8 financial technology personnel in a certain branch to test their satisfaction and practicality ratings before and after the optimization of the training system. Before optimizing the training system, the average satisfaction score of each technical personnel was 6.70 points, and the average practicality score was 5.59 points; After optimizing the training system, the average satisfaction score of each technical personnel was 8.56 points, and the average practicality score was 7.99 points. From the above data, it can be seen that after optimizing the personnel training system, technology personnel have higher satisfaction and practicality ratings for the training system, indicating that optimizing the personnel training system has certain practical significance.
- Open Access
- 1 January 2023
- Published by Association for Computational Linguistics (ACL)
- p. 611-615
YNU-HPCC at WASSA 2023: Using Text-Mixed Data Augmentation for Emotion Classification on Code-Mixed Text Message
Xuqiao Ran, You Zhang, Jin Wang, Dan Xu, Xuejie Zhang. Proceedings of the 13th Workshop on Computational Approaches to Subjectivity, Sentiment, & Social Media Analysis. 2023.
- 1 January 2023
- Published by Association for Computational Linguistics (ACL)
- p. 611-615
Open AccessYNU-HPCC at WASSA 2023: Using Text-Mixed Data Augmentation for Emotion Classification on Code-Mixed Text Message
Xuqiao Ran, You Zhang, Jin Wang, Dan Xu, Xuejie Zhang. Proceedings of the 13th Workshop on Computational Approaches to Subjectivity, Sentiment, & Social Media Analysis. 2023.