π Publications
π
2026
- Polymorphic Universal TransformerYilong Chen, Zitian Gao, Yihao Xiao, Jason Klein Liu, Xinyu Yang, Yifan Luo, 5 more authorsProceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), 2026
- LoopCoder: Scaling Code Intelligence via Looped Language ModelsJian Yang, Wei Zhang, Shuyue Guo, Yizhi Li, Linzheng Chai, Zhengmao Ye, 18 more authorsFindings of the Association for Computational Linguistics: ACL 2026
- DLRover-LM: LLM Pre-Training Framework with Thousands of Accelerators in AntGroupZiling Huang, Zhengmao Ye, Qingsong Cai, Zelong Huang, Bo Sang, Haitao Zhang, 5 more authors42nd IEEE International Conference on Data Engineering (ICDE), 2026
- Beyond N-gram: Data-Aware X-GRAM Extraction for Efficient Embedding Parameter ScalingYilong Chen, Yanxi Xie, Zitian Gao, He Xin, Yihao Xiao, Jason Klein Liu, 7 more authorsarXiv preprint arXiv:2604.21724, April 23, 2026
- Can You Trust the Vectors in Your Vector Database? Black-Hole Attack from Embedding Space DefectsHanxi Li, Jianan Zhou, Jiale Lao, Yibo Wang, Zhengmao Ye, Yang Cao, 2 more authorsarXiv preprint arXiv:2604.05480, April 7, 2026
- IQuest-Coder-V1 Technical ReportJian Yang, Wei Zhang, Shawn Guo, Zhengmao Ye, Lin Jing, Shark Liu, 32 more authorsarXiv preprint arXiv:2603.16733, March 17, 2026
π
2025
- Tuning Less, Prompting More: In-Context Preference Learning Pipeline for Natural Language TransformationShuyun Yang, Yan Zhang, Zhengmao Ye, Lei Duan, Mingjie TangProceedings of the Conference on Empirical Methods in Natural Language Processing (EMNLP), 2025
- mLoRA: Fine-Tuning LoRA Adapters via Highly-Efficient Pipeline Parallelism in Multiple GPUsZhengmao Ye, Dengchun Li, Zetao Hu, Tingfeng Lan, Jian Sha, Sicong Zhang, 5 more authorsProceedings of Very Large Data Bases Conference (VLDB), 2025
- Diffusion Counterfactual-Based Anomaly Detection in Class-Imbalanced DataXinyun Shen, Min Li, Zhengmao Ye, Zhenyang Yu, Lei DuanIEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP), 2025
π
2024
- MixLoRA: Enhancing Large Language Models Fine-Tuning with LoRA-based Mixture of ExpertsDengchun Li, Yingzi Ma, Naizheng Wang, Zhengmao Ye, Zhiyuan Cheng, Yinghao Tang, 5 more authorsarXiv preprint arXiv:2404.15159, 2024
π Projects
- An Efficient "Factory" to Build Multiple LoRA Adapters
- Collaborative Agent Workflow through Programming and Prompting
- Hyperledger Fabric 1.4.4, Hyperledger Caliper 0.3.0 configuration - pbft/solo/multichannel