Yuchang Sun

Yuchang Sun

Ph.D.

Publications

I work on post-training of large language models now. During my PhD study, I proposed frameworks/algorithms with theoretical guarantee to address the data and device heterogeneity in federated learning (FL). You can also find my articles on my Google Scholar profile.

Preprint Articles

Group-Relative REINFORCE Is Secretly an Off-Policy Algorithm: Demystifying Some Myths About GRPO and Its Friends

Published in Under review, 2025

Recommended citation: C. Yao, Y. Chen, Y. Sun, Y. Chen, W. Zhang, X. Pan, Y. Li, and B. Ding, “Group-Relative REINFORCE Is Secretly an Off-Policy Algorithm: Demystifying Some Myths About GRPO and Its Friends,” arxiv: 2509.24203.

On-Policy RL Meets Off-Policy Experts: Harmonizing Supervised Fine-Tuning and Reinforcement Learning via Dynamic Weighting

Published in Under review, 2025

Recommended citation: W. Zhang, Y. Xie, Y. Sun, Y. Chen, G. Wang, Y. Li, B. Ding, and J. Zhou, “On-Policy RL Meets Off-Policy Experts: Harmonizing Supervised Fine-Tuning and Reinforcement Learning via Dynamic Weighting,” arxiv: 2508.11408.

Enhancing Latent Computation in Transformers with Latent Tokens

Published in Under review, 2025

Recommended citation: Y. Sun, Y. Chen, Y. Li, and B. Ding, “Enhancing latent computation in transformers with latent tokens,” submitted.

Conference Papers

Learn How to Query from Unlabeled Data Streams in Federated Learning

Published in AAAI 2025, 2025

Recommended citation: Y. Sun, X. Li, T. Lin, and J. Zhang. (2024). "Learn How to Query from Unlabeled Data Streams in Federated Learning." AAAI 2025.
Paper | Introduction | Code

Exploring Selective Layer Fine-Tuning in Federated Learning

Published in IEEE ISIT 2025, 2024

Recommended citation: Y. Sun, Y. Xie, B. Ding, Y. Li, and J. Zhang, “Exploring selective layer fine-tuning in federated learning,” IEEE Int. Symp. Inf. Theory (ISIT), Michigan, USA, Jun. 2025.
Paper

DReS-FL: Dropout-Resilient Secure Federated Learning for Non-IID Clients via Secret Data Sharing

Published in NeurIPS 2022, 2022

Recommended citation: J. Shao, Y. Sun, S. Li, and J. Zhang. (2022). "DReS-FL: Dropout-Resilient Secure Federated Learning for Non-IID Clients via Secret Data Sharing." NeurIPS 2022.
Paper | Introduction | Code

Asynchronous Semi-Decentralized Federated Edge Learning for Heterogeneous Clients

Published in IEEE ICC 2022, 2022

Recommended citation: Y. Sun, J. Shao, Y. Mao, and J. Zhang, “Asynchronous semi-decentralized federated edge learning for heterogenous clients,” IEEE Int. Conf. Commun. (ICC), Seoul, South Korea, May 2022.
Paper

Stochastic Coded Federated Learning with Convergence and Privacy Guarantees

Published in IEEE ISIT 2022, 2022

Recommended citation: Y. Sun, J. Shao, S. Li, Y. Mao, and J. Zhang, “Stochastic coded federated learning with convergence and privacy guarantees,” IEEE Int. Symp. Inf. Theory (ISIT), Espoo, Finland, Jun.-Jul. 2022.
Paper | Introduction | Code

Semi-Decentralized Federated Edge Learning for Fast Convergence on Non-IID Data

Published in IEEE WCNC 2022, 2021

Recommended citation: Y. Sun, J. Shao, Y. Mao, J. Wang, and J. Zhang, “Semi-decentralized federated edge learning for fast convergence on non-IID data,” IEEE Wireless Commun. Networking Conf. (WCNC), Austin, TX, USA, Apr. 2022.
Paper

Journal Articles

How to Collaborate: Towards Maximizing the Generalization Performance in Cross-Silo Federated Learning

Published in IEEE Trans. Mobile Comput., 2024

Recommended citation: Y. Sun, M. Kountouris, and J. Zhang, “How to collaborate: Towards maximizing the generalization performance in cross-silo federated learning,” accepted to IEEE Trans. Mobile Comput.
Paper

Stochastic Coded Federated Learning: Theoretical Analysis and Incentive Mechanism Design

Published in IEEE Transactions on Wireless Communication, 2023

Recommended citation: Y. Sun, J. Shao, Y. Mao, S. Li, and J. Zhang, “Stochastic coded federated learning: The- oretical analysis and incentive mechanism design,” IEEE Trans. Wireless Commun., vol. 23, no. 6, pp. 6623-6638, Jun. 2024.
Paper

Channel and Gradient-Importance Aware Device Scheduling for Over-the-Air Federated Learning-the-Air Federated Learning

Published in IEEE Transactions on Wireless Communication on Wireless Communication, 2023

Recommended citation: Y. Sun, Z. Lin, Y. Mao, S. Jin, J. Zhang, “Channel and gradient-importance aware de- vice scheduling for over-the-air federated learning,” IEEE Trans. Wireless Commun., vol. 23, no. 7, pp. 6905-6920, Jul. 2024.
Paper

MimiC: Combating Client Dropouts in Federated Learning by Mimicking Central Updates

Published in IEEE Transactions on Mobile Computing (CCF A), 2023

Recommended citation: Y. Sun, Y. Mao, and J. Zhang, “MimiC: Combating client dropouts in federated learning by mimicking central updates,” IEEE Trans. Mob. Comput., to appear.
Paper | Introduction | Code

Semi-Decentralized Federated Edge Learning with Data and Device Heterogeneity

Published in IEEE Transactions on Network and Service Management, 2022

Recommended citation: Y. Sun, J. Shao, Y. Mao, J. Wang, and J. Zhang, “Semi-decentralized federated edge learning with data and device heterogeneity,” IEEE Trans. Netw. Service Manage., vol. 20, no. 2, pp. 1487-1501, Jun. 2023.
Paper | Introduction | Code