Cheng Yang
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Cheng Yang has a total of 40 co-authors. Here are the top co-authors by number of shared publications:
- Huaimin Wang
National University of Defense Technology
5 shared publications - Kele Xu
National university of defense technology, Changsha, China
5 shared publications - Bo Ding
National University of Defense Technology
4 shared publications - Dawei Feng
National University of Defense Technology
4 shared publications Gang Yin
National University of Defense Technology
2 shared publicationsQiang Fan
National University of Defense Technology
2 shared publications
Latest Publications
- Open Access
- 9 November 2026
- Published by Association for Computing Machinery (ACM)
- p. 29-33
HAP-Diff: Hierarchical Multimodal Diffusion with Personality Adaptation for Online Facial Reaction Generation
- 9 November 2026
- Published by Association for Computing Machinery (ACM)
- p. 29-33
Open AccessHAP-Diff: Hierarchical Multimodal Diffusion with Personality Adaptation for Online Facial Reaction Generation
- Open Access
- 13 April 2026
- Published by Association for Computing Machinery (ACM)
- p. 1-17
Divergence or Convergence? A Deep Insight into the Crowd Collaboration and its Productivity in Open Source Software based on Entropy
- 13 April 2026
- Published by Association for Computing Machinery (ACM)
- p. 1-17
Open AccessDivergence or Convergence? A Deep Insight into the Crowd Collaboration and its Productivity in Open Source Software based on Entropy
- Open Access
- 16 October 2025
- Published by Association for Computing Machinery (ACM) in Proceedings of the ACM on Human-Computer Interaction
- Vol. 9 (7) , 1-26
Are External Contributions Important to Project Productivity in Open Source Software? A Deep Insight based on Issue Entropy
In the realm of open source software (OSS) development, the resolution of issues is not just a technical task but a pivotal activity that drives ongoing enhancement and secures a project's sustainability. Contributing to issues is the majority form for external contributors to take part in OSS projects. Although the significance of external contributors is recognized, their contributions in issue process still lacks full quantification and clarity, and the correlation between their contributions and the project productivity remains unclear. In response, we propose issue entropy, a novel metric that quantifies the complexity of event sequence in issue process applied to study the external contributions. The metric applies principles of information theory to scrutinize granular details within a project's issue-related activities, providing a unique lens to understand and assess external contributions. To explore the correlation between external contributions and project productivity, we employed issue entropy as a novel way to examine external contributions, and analyzed its correlation with new bugs, commits, and bug fix time, serving as proxies for OSS project productivity. Our findings reveal a strong positive correlation with new bugs, variable relationship with commit volume based on company sponsorship, and significant negative correlation with bug fix time. We also observe the significant interactions between external contributions and other factors, such as project age and the number of files. Moreover, issue entropy offers a new perspective on the health of the OSS project ecosystem, potentially supporting further research and practical applications.
- 16 October 2025
- Published by Association for Computing Machinery (ACM) in Proceedings of the ACM on Human-Computer Interaction
- Vol. 9 (7) , 1-26
Open AccessAre External Contributions Important to Project Productivity in Open Source Software? A Deep Insight based on Issue Entropy
In the realm of open source software (OSS) development, the resolution of issues is not just a technical task but a pivotal activity that drives ongoing enhancement and secures a project's sustainability. Contributing to issues is the majority form for external contributors to take part in OSS projects. Although the significance of external contributors is recognized, their contributions in issue process still lacks full quantification and clarity, and the correlation between their contributions and the project productivity remains unclear. In response, we propose issue entropy, a novel metric that quantifies the complexity of event sequence in issue process applied to study the external contributions. The metric applies principles of information theory to scrutinize granular details within a project's issue-related activities, providing a unique lens to understand and assess external contributions. To explore the correlation between external contributions and project productivity, we employed issue entropy as a novel way to examine external contributions, and analyzed its correlation with new bugs, commits, and bug fix time, serving as proxies for OSS project productivity. Our findings reveal a strong positive correlation with new bugs, variable relationship with commit volume based on company sponsorship, and significant negative correlation with bug fix time. We also observe the significant interactions between external contributions and other factors, such as project age and the number of files. Moreover, issue entropy offers a new perspective on the health of the OSS project ecosystem, potentially supporting further research and practical applications.
- Open Access
- 13 July 2025
- Published by Association for Computing Machinery (ACM)
- p. 1444-1454
Empowering Large Language Model Agent through Step-Level Self-Critique and Self-Training
- 13 July 2025
- Published by Association for Computing Machinery (ACM)
- p. 1444-1454
Open AccessEmpowering Large Language Model Agent through Step-Level Self-Critique and Self-Training
- Open Access
- 13 July 2025
- Published by Association for Computing Machinery (ACM)
- p. 191-200
Preference-Strength-Aware Self-Improving Alignment with Generative Preference Models
- 13 July 2025
- Published by Association for Computing Machinery (ACM)
- p. 191-200
Open AccessPreference-Strength-Aware Self-Improving Alignment with Generative Preference Models
- Open Access
- 11 April 2025
- Published by Association for the Advancement of Artificial Intelligence (AAAI) in Proceedings of the AAAI Conference on Artificial Intelligence
- Vol. 39 (25) , 27161-27169
Enhancing Decision-Making for LLM Agents via Step-Level Q-Value Models
Agents significantly enhance the capabilities of standalone Large Language Models (LLMs) by perceiving environments, making decisions, and executing actions. However, LLM agents still face challenges in tasks that require multiple decision-making steps. Estimating the value of actions in specific tasks is difficult when intermediate actions are neither appropriately rewarded nor penalized. In this paper, we propose leveraging a task-relevant Q-value model to guide action selection. Specifically, we first collect decision-making trajectories annotated with step-level Q values via Monte Carlo Tree Search (MCTS) and construct preference data. We then use another LLM to fit these preferences through step-level Direct Policy Optimization (DPO), which serves as the Q-value model. During inference, at each decision-making step, LLM agents select the action with the highest Q value before interacting with the environment. We apply our method to various open-source and API-based LLM agents, demonstrating that Q-value models significantly improve their performance. Notably, the performance of the agent built with Phi-3-mini-4k-instruct improved by 103% on WebShop and 75% on HotPotQA when enhanced with Q-value models, even surpassing GPT-4o-mini. Additionally, Q-value models offer several advantages, such as generalization to different LLM agents and seamless integration with existing prompting strategies.
- 11 April 2025
- Published by Association for the Advancement of Artificial Intelligence (AAAI) in Proceedings of the AAAI Conference on Artificial Intelligence
- Vol. 39 (25) , 27161-27169
Open AccessEnhancing Decision-Making for LLM Agents via Step-Level Q-Value Models
Agents significantly enhance the capabilities of standalone Large Language Models (LLMs) by perceiving environments, making decisions, and executing actions. However, LLM agents still face challenges in tasks that require multiple decision-making steps. Estimating the value of actions in specific tasks is difficult when intermediate actions are neither appropriately rewarded nor penalized. In this paper, we propose leveraging a task-relevant Q-value model to guide action selection. Specifically, we first collect decision-making trajectories annotated with step-level Q values via Monte Carlo Tree Search (MCTS) and construct preference data. We then use another LLM to fit these preferences through step-level Direct Policy Optimization (DPO), which serves as the Q-value model. During inference, at each decision-making step, LLM agents select the action with the highest Q value before interacting with the environment. We apply our method to various open-source and API-based LLM agents, demonstrating that Q-value models significantly improve their performance. Notably, the performance of the agent built with Phi-3-mini-4k-instruct improved by 103% on WebShop and 75% on HotPotQA when enhanced with Q-value models, even surpassing GPT-4o-mini. Additionally, Q-value models offer several advantages, such as generalization to different LLM agents and seamless integration with existing prompting strategies.
- Open Access
- 28 October 2024
- Published by Association for Computing Machinery (ACM)
- p. 11457-11463
Higher-Order Vision-Language Alignment for Social Media Prediction
The prediction task of social media popularity aims to automatically forecast the future popularity of the posts by leveraging vast amounts of social media data. This data encompasses diverse visual and textual content, including photos, categories, custom tags, temporal information, and geographical data. Existing methods have explored multiple feature types to enhance popularity prediction. Despite their success, visual and textual features-both crucial pieces of information-are often simply concatenated after extraction, ignoring the divergence between these two feature spaces. In this paper, we propose a method to project visual and language information into an aligned semantic representation, thereby uncovering intricate associations between these two modalities. Specifically, we leverage the BLIP-2 model to understand and generate visual description text that encapsulates the content of photos. Semantic embeddings are then extracted from all available visual and textual information. Additionally, we deeply exploit user-related behavior and characteristic information to extract features, uncovering hidden clues for post popularity prediction. Leveraging these improvements, we conduct extensive experiments to demonstrate the effectiveness of our proposed method.
- 28 October 2024
- Published by Association for Computing Machinery (ACM)
- p. 11457-11463
Open AccessHigher-Order Vision-Language Alignment for Social Media Prediction
The prediction task of social media popularity aims to automatically forecast the future popularity of the posts by leveraging vast amounts of social media data. This data encompasses diverse visual and textual content, including photos, categories, custom tags, temporal information, and geographical data. Existing methods have explored multiple feature types to enhance popularity prediction. Despite their success, visual and textual features-both crucial pieces of information-are often simply concatenated after extraction, ignoring the divergence between these two feature spaces. In this paper, we propose a method to project visual and language information into an aligned semantic representation, thereby uncovering intricate associations between these two modalities. Specifically, we leverage the BLIP-2 model to understand and generate visual description text that encapsulates the content of photos. Semantic embeddings are then extracted from all available visual and textual information. Additionally, we deeply exploit user-related behavior and characteristic information to extract features, uncovering hidden clues for post popularity prediction. Leveraging these improvements, we conduct extensive experiments to demonstrate the effectiveness of our proposed method.
- Open Access
- 5 July 2024
- Published by Association for Computing Machinery (ACM)
- p. 101-107
A Physically Inspired Graph Neural Network for Crystal Properties Prediction
Crystal is a kind of material composed of unit cells arranged in a periodic pattern in three-dimensional space. Predicting properties is a crucial task in crystal materials research. Accurate representation of the structure and precise property prediction are essential for this task. Graph-based representations of structures have been widely adopted to model atoms and bonds, and significant advancements have been made using graph learning methods. In this work, we propose a novel approach that leverages physical knowledge to enhance the modeling of interatomic relationships. Specifically, we replace the traditional Euclidean distance with interatomic potential energy and incorporate approximate kinetic energy into the features of atomic nodes. These enhanced representations are then integrated into our proposed neural network, named PBCGNN. Our method is experimentally evaluated on the JARVIS datasets, demonstrating that this physics-inspired approach yields improved performance compared to multiple existing methods.
- 5 July 2024
- Published by Association for Computing Machinery (ACM)
- p. 101-107
Open AccessA Physically Inspired Graph Neural Network for Crystal Properties Prediction
Crystal is a kind of material composed of unit cells arranged in a periodic pattern in three-dimensional space. Predicting properties is a crucial task in crystal materials research. Accurate representation of the structure and precise property prediction are essential for this task. Graph-based representations of structures have been widely adopted to model atoms and bonds, and significant advancements have been made using graph learning methods. In this work, we propose a novel approach that leverages physical knowledge to enhance the modeling of interatomic relationships. Specifically, we replace the traditional Euclidean distance with interatomic potential energy and incorporate approximate kinetic energy into the features of atomic nodes. These enhanced representations are then integrated into our proposed neural network, named PBCGNN. Our method is experimentally evaluated on the JARVIS datasets, demonstrating that this physics-inspired approach yields improved performance compared to multiple existing methods.
- Open Access
- 12 March 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Open Journal of Engineering in Medicine and Biology
- Vol. 5, 226-237
Masked Modeling-Based Ultrasound Image Classification via Self-Supervised Learning
Recently, deep learning-based methods have emerged as the preferred approach for ultrasound data analysis. However, these methods often require large-scale annotated datasets for training deep models, which are not readily available in practical scenarios. Additionally, the presence of speckle noise and other imaging artifacts can introduce numerous hard examples for ultrasound data classification. In this paper, drawing inspiration from self-supervised learning techniques, we present a pre-training method based on mask modeling specifically designed for ultrasound data. Our study investigates three different mask modeling strategies: random masking, vertical masking, and horizontal masking. By employing these strategies, our pre-training approach aims to predict the masked portion of the ultrasound images. Notably, our method does not rely on externally labeled data, allowing us to extract representative features without the need for human annotation. Consequently, we can leverage unlabeled datasets for pre-training. Furthermore, to address the challenges posed by hard samples in ultrasound data, we propose a novel hard sample mining strategy. To evaluate the effectiveness of our proposed method, we conduct experiments on two datasets. The experimental results demonstrate that our approach outperforms other state-of-the-art methods in ultrasound image classification. This indicates the superiority of our pre-training method and its ability to extract discriminative features from ultrasound data, even in the presence of hard examples.
- 12 March 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Open Journal of Engineering in Medicine and Biology
- Vol. 5, 226-237
Open AccessMasked Modeling-Based Ultrasound Image Classification via Self-Supervised Learning
Recently, deep learning-based methods have emerged as the preferred approach for ultrasound data analysis. However, these methods often require large-scale annotated datasets for training deep models, which are not readily available in practical scenarios. Additionally, the presence of speckle noise and other imaging artifacts can introduce numerous hard examples for ultrasound data classification. In this paper, drawing inspiration from self-supervised learning techniques, we present a pre-training method based on mask modeling specifically designed for ultrasound data. Our study investigates three different mask modeling strategies: random masking, vertical masking, and horizontal masking. By employing these strategies, our pre-training approach aims to predict the masked portion of the ultrasound images. Notably, our method does not rely on externally labeled data, allowing us to extract representative features without the need for human annotation. Consequently, we can leverage unlabeled datasets for pre-training. Furthermore, to address the challenges posed by hard samples in ultrasound data, we propose a novel hard sample mining strategy. To evaluate the effectiveness of our proposed method, we conduct experiments on two datasets. The experimental results demonstrate that our approach outperforms other state-of-the-art methods in ultrasound image classification. This indicates the superiority of our pre-training method and its ability to extract discriminative features from ultrasound data, even in the presence of hard examples.
- Open Access
- 1 January 2021
- Published by Wiley in Wireless Communications and Mobile Computing
- Vol. 2021 (1)
TagNN: A Code Tag Generation Technology for Resource Retrieval from Open‐Source Big Data
With the vigorous development of open‐source software, a huge number of open‐source projects and open‐source codes have been accumulated in open‐source big data, which contains a wealth of code resources. However, effectively and efficiently retrieving the relevant code snippets in such a large amount of open‐source big data is an extremely difficult problem. There are usually large gaps between the user’s natural language description and the open‐source code snippets. In this paper, we propose a novel code tag generation and code retrieval approach named TagNN, which combines software engineering empirical knowledge and a deep learning algorithm. The experimental results show that our method has good effects on code tag generation and code snippet retrieval.
- 1 January 2021
- Published by Wiley in Wireless Communications and Mobile Computing
- Vol. 2021 (1)
Open AccessTagNN: A Code Tag Generation Technology for Resource Retrieval from Open‐Source Big Data
With the vigorous development of open‐source software, a huge number of open‐source projects and open‐source codes have been accumulated in open‐source big data, which contains a wealth of code resources. However, effectively and efficiently retrieving the relevant code snippets in such a large amount of open‐source big data is an extremely difficult problem. There are usually large gaps between the user’s natural language description and the open‐source code snippets. In this paper, we propose a novel code tag generation and code retrieval approach named TagNN, which combines software engineering empirical knowledge and a deep learning algorithm. The experimental results show that our method has good effects on code tag generation and code snippet retrieval.