Back to Zeyi Wen's Homepage; lead author(s) with ^# denote my advised students at the time.

2026

• Tong Yuan^#, Chengxi Liao^#, and Zeyi Wen.
  Strong Drafts Need Compact Memories: Long-Context Speculative Decoding with Compressed KV Cache.
  The 2026 Conference on Empirical Methods in Natural Language Processing (EMNLP), Findings, 2026, to appear.

• Yuebin Xu^#, Xuemei Peng^#, Junlan Chen^#, Zhiyi Chen^#, and Zeyi Wen.
  TRACE: Single-Pass Decoding-Trace Risk Localization for Generation Calibration.
  The 2026 Conference on Empirical Methods in Natural Language Processing (EMNLP), Findings, 2026, to appear.

• Yile Chen^#, Zeyuan Lin^#, Yawen Chen, Jian Chen, Zeyi Wen, and Jin Huang.
  Mining Input-Inherent Context: A Parameter-Free Kernel Attention Mechanism for CNNs.
  The IEEE International Conference on Data Mining (ICDM), 2026, to appear.

• Jian Chen, Yile Chen, Zhenya Zheng, Zeyi Wen, Yawen Chen, and Jin Huang.
  EaSFE: Scalable and Efficient Feature Engineering for Boosting Machine Learning Performance.
  ACM Transactions on Knowledge Discovery from Data (TKDD), accepted in May 2026.

• Xinhao Huang^#, Youliang Huang^#, and Zeyi Wen.
  Representation Drift Compensation: A Zero-Cost Enhancement for LLM Decomposition.
  The 43rd International Conference on Machine Learning (ICML), 2026, to appear.

• Minping Chen^#, Bing Xu, Zulong Chen, Chuanfei Xu, Ying Zhou, Zui Tao, and Zeyi Wen.
  Enhancing Online Recruitment with Category-Aware MoE and LLM-based Data Augmentation.
  The 64th Annual Meeting of the Association for Computational Linguistics (ACL), industry track, 2026, to appear.

• Minping Chen^#, Bowen Xiao^#, Liang Du, Chuxuan Zeng, and Zeyi Wen.
  Efficient Hyperparameter Optimization for LLM Reinforcement Learning.
  The 64th Annual Meeting of the Association for Computational Linguistics (ACL), main track, 2026, to appear.

• Ruijia Yang^#, and Zeyi Wen.
  An Efficient Heterogeneous Co-Design for Fine-Tuning on a Single GPU.
  The 63rd DAC, The Chips to Systems Conference (DAC), 2026, to appear.

• Youliang Huang^#, Xinhao Huang^#, Chengxi Liao^#, and Zeyi Wen.
  DeInfer: Efficient Parallel Inferencing for Decomposed Large Language Models.
  The 63rd DAC, The Chips to Systems Conference (DAC), 2026, to appear.

• Chengxi Liao^#, and Zeyi Wen.
  Channel-Aware Mixed-Precision Quantization for Efficient Long-Context Inference.
  The Fourteenth International Conference on Learning Representations (ICLR), 2026, to appear. PDF

• Yile Chen^#, Zeyi Wen, Jian Chen, and Jin Huang.
  Towards Dynamic Interleaving Optimizers.
  The Fourteenth International Conference on Learning Representations (ICLR), 2026, to appear. PDF

• Yuebin Xu^#, Yanhong Wang^#, Xuemei Peng^#, Hui Zang, Minghao Chen, Pengfei Xia, and Zeyi Wen.
  TD-MoE: Tensor Decomposition for MoE Models.
  The Fourteenth International Conference on Learning Representations (ICLR), 2026, to appear. PDF

• Rui Zhang^#, Boxuan Yang^#, Rongji Wang^#, Xuemei Peng^#, and Zeyi Wen.
  Efficient MoE Inference on Single Consumer-grade GPU with Dynamic Expert Caching.
  The 40th IEEE International Parallel and Distributed Processing Symposium (IPDPS), 2026, to appear.

• Jihang Li^#, Qing Liu, Zulong Chen, Jing Wang, Wei Wang, Chuanfei Xu, and Zeyi Wen.
  Taxon: Hierarchical Tax Code Prediction with Semantically Aligned LLM Expert Guidance.
  The 42nd IEEE International Conference on Data Engineering (ICDE), industry and application track, 2026, to appear. PDF

• Yile Chen^#, Zeyi Wen, Jian Chen, and Jin Huang.
  A Better Start: Sensitivity-Aware Warm-Up for Robust and Efficient Fine-Tuning.
  The 40th AAAI Conference on Artificial Intelligence (AAAI), 2026, to appear. PDF

2025

• JihangLi^#, Bing Xu, Zulong Chen, Chuanfei Xu, Minping Chen, Suyu Liu, Ying Zhou, and Zeyi Wen.
  Enhancing Talent Search Ranking with Role-Aware Expert Mixtures and LLM-based Fine-Grained Job Descriptions.
  The 2025 Conference on Empirical Methods in Natural Language Processing (EMNLP), industry track, 2025, to appear. PDF

• Yuebin Xu^#, Zhiyi Chen^#, and Zeyi Wen.
  EcoTune: Token-Efficient Multi-Fidelity Hyperparameter Optimization for Large Language Model Inference.
  The 2025 Conference on Empirical Methods in Natural Language Processing (EMNLP), 2025, to appear. PDF

• Xinhao Huang^#, Zhibo Ren, Yipeng Yu, Ying Zhou, Zulong Chen, and Zeyi Wen.
  SEAL: Structure and Element Aware Learning Improves Long Structured Document Retrieval.
  The 2025 Conference on Empirical Methods in Natural Language Processing (EMNLP), 2025, to appear. PDF

• Minping Chen^#, Ruijia Yang^#, and Zeyi Wen.
  Efficient Mask Learning for Language Model Fine-Tuning.
  The 34th Conference on Information and Knowledge Management (CIKM), 2025, to appear. PDF

• Jihang Li^#, Jiacheng Qiu, Yin-Ping Zhao, and Zeyi Wen.
  Tilia: Enhancing LIME with Decision Tree Surrogates.
  The 34th Conference on Information and Knowledge Management (CIKM), 2025, to appear. PDF

• Yuebin Xu^#, Xuemei Peng^#, Zhiyi Chen^#, and Zeyi Wen.
  KALE: Knowledge Aggregation for Label-free Model Enhancement.
  The 34th Conference on Information and Knowledge Management (CIKM), 2025, to appear. PDF

• Yawen Chen, Zeyi Wen, Jian Chen, and Jin Huang.
  Leveraging Pre-trained Models for Kernel Machines.
  Pattern Recognition (PR), accepted in June 2025.

• Hanfeng Liu^#, Xuemei Peng^#, and Zeyi Wen.
  Accelerating Multi-Output GBDTs with GPUs.
  The 54th International Conference on Parallel Processing (ICPP), 2025. PDF

• Jian Chen, Yile Chen, Zeyi Wen, Yawen Chen, and Jin Huang.
  Towards Recommendation on Good Quality Data Science Solutions.
  ACM Transactions on Knowledge Discovery from Data (TKDD), accepted in June 2025.

• Borui Xu, Yao Chen, Zeyi Wen, Weiguo Liu, and Bingsheng He.
  Evaluating Small Language Models for News Summarization: Implications and Factors Influencing Performance.
  Annual Conference of the Nations of the Americas Chapter of the ACL (NAACL), 2025, to appear. PDF

• Minping Chen^#, You-Liang Huang^#, and Zeyi Wen.
  Towards Efficient Low-order Hybrid Optimizer for Language Model Fine-tuning.
  The 39th Annual AAAI Conference on Artificial Intelligence (AAAI), 2025. PDF

• Xinhao Huang^#, You-Liang Huang^#, and Zeyi Wen.
  SoLA: Leveraging Soft Activation Sparsity and Low-Rank Decomposition for Large Language Model Compression.
  The 39th Annual AAAI Conference on Artificial Intelligence (AAAI), 2025. PDF

• Borui Xu, Zeyi Wen, Yao Chen, Weiguo Liu, Weng-Fai Wong, and Bingsheng He.
  ScalaGBM: Memory Efficient GBDT Training for High-Dimensional Data on GPU.
  The 31st SIGKDD Conference on Knowledge Discovery and Data Mining (KDD), 2025, to appear. PDF

2024

• Hanfeng Liu^#, Minping Chen^#, Zhenya Zheng^#, and Zeyi Wen.
  Exploiting Careful Design of SVM Solution for Aspect-term Sentiment Analysis.
  The 2024 Conference on Empirical Methods in Natural Language Processing (EMNLP), Findings, pages 5897-5906. PDF

• Jiantong Jiang^#, Zeyi Wen, Atif Mansoor, and Ajmal Saeed Mian.
  Faster-BNI: Fast Parallel Exact Inference on Bayesian Networks.
  IEEE Transactions on Parallel and Distributed Systems (TPDS), accepted in Jun 2024. PDF

• Jiantong Jiang^#, Zeyi Wen, Atif Mansoor, and Ajmal Saeed Mian.
  Fast Inference for Probabilistic Graphical Models.
  The 2024 USENIX Annual Technical Conference (USENIX ATC), to appear. PDF

• Jiantong Jiang^#, Zeyi Wen, Atif Mansoor, and Ajmal Saeed Mian.
  Efficient Hyperparameter Optimization with Adaptive Fidelity Identification.
  The IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), to appear 2024. PDF

• Yile Chen^#, Zeyi Wen, Jian Chen, and Jin Huang.
  Enhancing the Performance of Bandit-based Hyperparameter Optimization.
  The 40^th IEEE International Conference on Data Engineering (ICDE), to appear 2024. PDF

2023

• Jian Yang^#, Jiantong Jiang^#, Zeyi Wen, and Ajmal Mian.
  Parallel and Distributed Bayesian Network Structure Learning.
  IEEE Transactions on Parallel and Distributed Systems (TPDS), accepted in Oct 2023. PDF

• Borui Xu^#, Zeyi Wen, Lifeng Yan, Zhan Zhao, Zekun Yin, Weiguo Liu and Bingsheng He.
  Leveraging Data Density and Sparsity for Efficient SVM Training on GPUs.
  The 23^rd IEEE International Conference on Data Mining (ICDM), 2023. PDF

• Zeyi Wen, Mingyu Liang, Bingsheng He and Zexin Xia.
  A High-Performance Index for Real-Time Matrix Retrieval.
  The 39^th IEEE International Conference on Data Engineering (ICDE), 2023 (poster)

• Peiyu Yang, Naveed Akhtar, Zeyi Wen, Mubarak Shah and Ajmal Mian.
  Re-calibrating Feature Attributions for Model Interpretation.
  The International Conference on Learning Representations (ICLR), 2023. PDF

• Peiyu Yang, Naveed Akhtar, Zeyi Wen and Ajmal Mian.
  Local Path Integration for Attribution.
  The 37^th AAAI Conference on Artificial Intelligence (AAAI), 2023. PDF

• Jiantong Jiang^#, Zeyi Wen, Atif Mansoor and Ajmal Mian.
  Fast Parallel Exact Inference on Bayesian Networks.
  The 28^th ACM SIGPLAN Annual Symposium on Principles and Practice of Parallel Programming (PPoPP), 2023 (poster). PDF

2022

• Yawen Chen^#, Yile Chen^#, Jian Chen, Zeyi Wen and Jin Huang.
  Efficient Second-Order Optimization for Neural Networks with Kernel Machines.
  The 30^th ACM International Conference on Information and Knowledge Management (CIKM), 2022. PDF

• Jiantong Jiang^#, Zeyi Wen and Ajmal Mian.
  Fast Parallel Bayesian Network Structure Learning.
  The 36^th IEEE International Parallel and Distributed Processing Symposium (IPDPS), pages 617-627, 2022. PDF

• Jiantong Jiang^#, Zeyi Wen, Zeke Wang, Bingsheng He and Jian Chen.
  Parallel and Distributed Structured SVM Training.
  IEEE Transactions on Parallel and Distributed Systems (TPDS), 33(5):1084-1096, 2022. PDF

• Peiyu Yang^#, Zeyi Wen, Ajmal Mian.
  Multi-Grained Interpretable Network for Image Recognition.
  The 26^th International Conference on Pattern Recognition (ICPR), 2022. PDF

2021

• Yawen Chen^#, Zeyi Wen, Bingsheng He and Jian Chen.
  Efficient Decomposition Selection for Multi-class Classification.
  IEEE Transactions on Knowledge and Data Engineering (TKDE), 35(4):3751 - 3764, 2021. PDF

• Qinbin Li, Zeyi Wen, Zhaomin Wu, Sixu Hu, Naibo Wang, Yuan Li, Xu Liu and Bingsheng He.
  A Survey on Federated Learning Systems: Vision, Hype and Reality for Data Privacy and Protection.
  IEEE Transactions on Knowledge and Data Engineering (TKDE), 35(4):3347 - 3366, 2021. PDF

• Hanfeng Liu^#, Zeyi Wen and Wei Cai.
  FastPSO: Towards Efficient Swarm Intelligence Algorithm on GPUs.
  The 50^th International Conference on Parallel Processing (ICPP), pages 1-10, 2021. PDF (Best Paper Award Finalist: 4 out of 329)

• Zeyi Wen, Zhishang Zhou, Hanfeng Liu, Bingsheng He, Xia Li and Jian Chen.
  Enhancing SVMs with Problem Context Aware Pipeline.
  The 27^th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (KDD), Research Track, pages 1821-1829, 2021. PDF

• Zeyi Wen, Qinbin Li, Bingsheng He and Bin Cui.
  Challenges and Opportunities of Building Fast GBDT Systems.
  The 30^th International Joint Conference on Artificial Intelligence (IJCAI), Survey Track, pages 4661-4668, 2021. PDF

2020

• Zeyi Wen, Mingyu Liang, Bingsheng He and Zexin Xia.
  A High-Performance Index for Real-Time Matrix Retrieval.
  IEEE Transactions on Knowledge and Data Engineering (TKDE), 34(7): 3044-3056, 2022. (Accepted in Sep 2020) PDF

• Zeyi Wen, Hanfeng Liu, Jiashuai Shi, Qinbin Li, Bingsheng He and Jian Chen.
  ThunderGBM: Fast GBDTs and Random Forests on GPUs.
  Journal of Machine Learning Research (JMLR), 21(1):4389-4393, 2020. PDF

• Qinbin Li, Zhaomin Wu, Zeyi Wen and Bingsheng He.
  Privacy-Preserving Gradient Boosting Decision Trees.
  The 34^thAAAI Conference on Artificial Intelligence (AAAI), pages 784-791, 2020. PDF

• Qinbin Li, Zeyi Wen and Bingsheng He.
  Practical Federated Gradient Boosting Decision Trees.
  The 34^th AAAI Conference on Artificial Intelligence (AAAI), pages 4642-4649, 2020. PDF

• Zeyi Wen, Zeyu Huang and Rui Zhang.
  Entity Extraction with Knowledge from Web Scale Corpora.
  The 31^st Australasian Database Conference (ADC), pages 173-185, 2020. PDF

2019

• Qinbin Li, Zeyi Wen and Bingsheng He.
  Adaptive Kernel Value Caching for SVM Training.
  IEEE Transactions on Neural Networks and Learning Systems (TNNLS), 31(7):2376-2386, 2019. PDF

• Zeyi Wen, Jiashuai Shi, Bingsheng He, Jian Chen, Kotagiri Ramamohanarao and Qinbin Li.
  Exploiting GPUs for Efficient Gradient Boosting Decision Tree Training.
  IEEE Transactions on Parallel and Distributed Systems (TPDS), 30(12):2706-2717, 2019. PDF (Best Paper Award: 1 out of 987)

• Zeyi Wen, Dong Deng, Rui Zhang, and Kotagiri Ramamohanarao.
  An Efficient Entity Extraction Algorithm using Two-Level Edit-Distance.
  The 35^th IEEE International Conference on Data Engineering (ICDE), pages 998-1009, 2019. PDF

• Zeyi Wen, Mingyu Liang, Bingsheng He, Zexin Xia, and Bo Li.
  Aucher: Multi-modal Queries on Live Audio Streams in Real-time.
  The 35^th IEEE International Conference on Data Engineering (ICDE), pages 1960-1963, 2019 (demo). PDF

• Zeyi Wen, Jiashuai Shi, Bingsheng He, Jian Chen, and Yawen Chen.
  Efficient Multi-Class Probabilistic SVMs on GPUs.
  The 35^th IEEE International Conference on Data Engineering (ICDE), pages 2119-2120, 2019 (poster). PDF



2018

• Zeyi Wen, Jiashuai Shi, Bingsheng He, Jian Chen, and Yawen Chen.
  Efficient Multi-Class Probabilistic SVMs on GPUs.
  IEEE Transactions on Knowledge and Data Engineering (TKDE), pages 1693-1706, 2018. PDF

• Zeyi Wen, Jiashuai Shi, Qinbin Li, Bingsheng He, and Jian Chen.
  ThunderSVM: A Fast SVM Library on GPUs and CPUs.
  Journal of Machine Learning Research (JMLR), volume 19, pages 797-801, 2018. PDF

• Zeyi Wen, Bingsheng He, Kotagiri Ramamohanarao, Shengliang Lu, and Jiashuai Shi.
  Efficient Gradient Boosted Decision Tree Training on GPUs.
  The 32^nd IEEE International Parallel and Distributed Processing Symposium (IPDPS), pages 234-243, 2018. PDF

• Zeyi Wen, Xingyang Liu, Hongjian Cao, and Bingsheng He.
  RTSI: An Index Structure for Multi-modal Real-Time Search on Live Audio Streaming Services.
  The 34^th IEEE International Conference on Data Engineering (ICDE), pages 1495-1506, 2018. PDF

• Jiashuai Shi, Zeyi Wen, Bingsheng He, and Jian Chen.
  Efficient Support Vector Machine Training Algorithm on GPUs.
  The 32^nd AAAI Conference on Artificial Intelligence (AAAI), pages 8157-8158, 2018. PDF

• Yawen Chen, Zeyi Wen, Jian Chen and Jin Huang.
  Selecting Proper Multi-class SVM Training Methods.
  The 32^nd AAAI Conference on Artificial Intelligence (AAAI), pages 8061-8062, 2018. PDF



2017 and before

• Zeyi Wen, Bin Li, Kotagiri Ramamohanarao, Jian Chen, Yawen Chen and Rui Zhang.
  Improving Efficiency of SVM k-Fold Cross-Validation by Alpha Seeding.
  The 31^st AAAI Conference on Artificial Intelligence (AAAI), pages 2768-2774, 2017. PDF

• Zeyi Wen, Rui Zhang, Kotagiri Ramamohanarao, and Li Yang.
  Scalable and Fast SVM Regression using Modern Hardware.
  World Wide Web Journal (WWWJ), 21(2):261-287, 2018. (accepted in Feb 2017) PDF

• Jian Chen, Jin Huang, Zeyi Wen, Zhen He, Kerry Taylor and Rui Zhang.
  Analysis and Evaluation of the top-k Most Influential Location Selection Query.
  Knowledge and Information Systems (KAIS), 43(1):181-217, 2015. PDF

• Jianzhong Qi, Rui Zhang, Kotagiri Ramamohanarao, Hongzhi Wang, Zeyi Wen and Dan Wu.
  Indexable Online Time Series Segmentation with Error Bound Guarantee.
  World Wide Web Journal (WWWJ), 18(2): 359-401, 2015. PDF

• Saad Aljubayrin, Jianzhong Qi, Christian Jensen, Rui Zhang, Zhen He, and Zeyi Wen.
  The Safest Path via Safe Zones.
  The 31^st IEEE International Conference on Data Engineering (ICDE), pages 531-542, 2015. PDF

• Zeyi Wen, Rui Zhang, Kotagiri Ramamohanarao, Jianzhong Qi and Kerry Taylor.
  MASCOT: Fast and Highly Scalable SVM Cross-validation using GPUs and SSDs.
  The IEEE International Conference on Data Mining (ICDM), pages 580-589, 2014. PDF

• Zeyi Wen, Rui Zhang and Kotagiri Ramamohanarao.
  Enabling Precision/Recall Preferences for Semi-supervised SVM Training.
  The 23^rd ACM Conference on Information and Knowledge Management (CIKM), pages 421-430, 2014. PDF

• Xiaojie Lin, Rui Zhang, Zeyi Wen, Hongzhi Wang and Jianzhong Qi.
  Efficient Subgraph Matching using GPUs.
  The 25^th Australasian Database Conference (ADC), pages 74-85, 2014. PDF

• Jianzhong Qi, Zhenghua Xu, Yuan Xue and Zeyi Wen.
  A Branch and Bound Method for Min-dist Location Selection Queries.
  The 23^rd Australasian Database Conference (ADC), pages 51-60, 2012. PDF

• Jin Huang, Zeyi Wen, Mukaddim Pathan, Kerry Taylor, Yuan Xue and Rui Zhang.
  Ranking Locations for Facility Selection based on Potential Influences.
  The 37^th Annual Conference of the IEEE Industrial Electronics Society (IECON), pages 2411-2416, 2011. PDF

• Jin Huang, Zeyi Wen, Jianzhong Qi, Rui Zhang, Jian Chen and Zhen He.
  Top-k Most Influential Location Selection.
  The 20^th ACM Conference on Information and Knowledge Management (CIKM), pages 2377-2380, 2011. PDF