zhanghainan

Hainan Zhang的个人主页

View My GitHub Profile

个人介绍 Introduction

张海楠于2019年获得中国科学院计算技术研究所博士学位。博士毕业后，就职于京东推荐研发部，任算法工程师。2023年加入北京航空航天大学人工智能学院未来区块链与隐私计算高精尖创新中心，任副研究员。目前主要研究方向为：自然语言生成、隐私计算、代码生成、对话系统等。在包括ACL、SIGIR、AAAI、EMNLP等国际学术会议和期刊上，以第一作者和通讯作者发表论文20余篇(其中CCF A类论文9篇, CCF B类论文7篇)，申请国家发明专利10余项，并获得AAAI-IAAI2022人工智能创新应用奖、中国科学院院长优秀奖、中科院计算所所长特别奖。受邀担任ICML、NIPS、TKDE、ACL、AAAI、SIGIR等会议审稿人。

Hainan Zhang, received her Ph.D. from the Institute of Computing Technology, Chinese Academy of Sciences in 2019. After graduation, she worked as an algorithm engineer in the R&D department, JD.com. In 2023, she joined the Future Blockchain and Privacy Computing Advanced Innovation Center of the Artificial Intelligence Research Institute of Beijing University of Aeronautics and Astronautics as an Associate Researcher. Her main research directions are: natural language generation, privacy computing, code generation, dialogue generation, etc. In recent years, she has published 20+ papers as the first author and corresponding author in NLP top conferences and journals, including ACL, SIGIR, AAAI, EMNLP. What’more, she also applied 10 national invention patents, the excellence award of the president of the Chinese Academy of Sciences and the special award of the director of the Institute of Computing and Technology. She was invited to serve as a PC for ICML、NIPS、TKDE、ACL、AAAI、SIGIR and other conferences.

招本科实习生

深度参与前沿研究，欢迎有保研需求或者博士申请的同学联系我们 ~~ zhanghainan@buaa.edu.cn

近期学术News

Best Paper Award(WWW2025 workshop)：HSF: Defending against Jailbreak Attacks with Hidden State Filtering

AAAI2025 Main收录3篇：

(1) MaFeRw: Query Rewriting with Multi-Aspect Feedbacks for Retrieval-Augmented Large Language Models

(2) Defending Against Sophisticated Poisoning Attacks with RL-based Aggregation in Federated Learning

(3) High-Fidelity Polarimetric Implicit 3D Reconstruction with View-Dependent Physical Representation

EMNLP2024 Main收录：Safely Learning with Private Data: A Federated Learning Framework for Large Language Model

发表论文 Publications

  1. (Best Paper) Cheng Qian, Hainan Zhang#, Lei Sha. HSF: Defending against Jailbreak Attacks with Hidden State Filtering.[C]//WWW2025 Workshop.
  2. (CCF A) Yujing Wang, Hainan Zhang#, Liang Pang, Hongwei Zheng, Zhiming Zheng. MaFeRw: Query Rewriting with Multi-Aspect Feedbacks for Retrieval-Augmented Large Language Models.[C]//AAAI2025.
  3. (CCF A) Yujing Wang, Hainan Zhang#, Sijia Wen, Wangjie Qiu, Binghui Guo. Defending Against Sophisticated Poisoning Attacks with RL-based Aggregation in Federated Learning.[C]//AAAI2025.
  4. (CCF A) Yu Qiu, Sijia Wen, Hainan Zhang, Zhiming Zheng. High-Fidelity Polarimetric Implicit 3D Reconstruction with View-Dependent Physical Representation. [C]//AAAI2025.
  5. (CCF B) JiaYing Zheng, HaiNan Zhang#, LingXiang Wang, WangJie Qiu, HongWei Zheng, ZhiMing Zheng. Safely Learning with Private Data: A Federated Learning Framework for Large Language Model.[C]//EMNLP2024.
  6. (CCF B) Xu Wang, Hainan Zhang#, Shuai Zhao, Hongshen Chen, Zhuoye Ding, Zhiguo Wan, Bo Cheng, and Yanyan Lan. Debiasing Counterfactual Context with Causal Inference for Multi-turn Dialogue Reasoning[J]. IEEE Transactions on Audio, Speech and Language Processing, 2023.
  7. (CCF B) Zhiping Wang, Pen Lin, Hainan Zhang#, Hongshen Chen, Tianhao Li, Zhuoye Ding, Sulong Xu and Jinghe Hu. Automatic Marketing Theme and Commodity Construction System for E-commerce.// EMNLP2023.
  8. (CCF C) Xu Wang, Hainan Zhang#, Shuai Zhao, Hongshen Chen, Bo Cheng, Zhuoye Ding, Sulong Xu, Weipeng Yan, Yanyan Lan. HiBERT: Detecting the illogical patterns with hierarchical BERT for multi-turn dialogue reasoning[J]. Neurocomputing, 2023, 524: 167-177.
  9. (CCF B) Yue Fang, Hainan Zhang#, Hongshen Chen, Zhuoye Ding, Bo Long, Yanyan Lan, Yanquan Zhou. From spoken dialogue to formal summary: An utterance rewriting for dialogue summarization.// NAACL2022.
  10. (CCF A) Xueying Zhang, Yanyan Zou, Hainan Zhang, Jing Zhou, Shiliang Diao, Jiajia Chen, Zhuoye Ding, Zhen He, Xueqi He, Yun Xiao, Bo Long, Han Yu, Lingfei Wu. Automatic Product Copywriting for E-Commerce[C].// AAAI2022.
  11. (CCF B) Haoran Xu, Hainan Zhang#, Yanyan Zou, Hongshen Chen, Zhuoye Ding and Yanyan Lan. Adaptive Bridge between Training and Inference for Dialogue Generation.[C]//EMNLP 2021.
  12. (CCF B) Xu Wang, Hainan Zhang#, Shuai Zhao, Yanyan Zou, Hongshen Chen, Zhuoye Ding, Bo Cheng and Yanyan Lan. FCM: A Fine-grained Comparison Model for Multi-turn Dialogue Reasoning.[C]//EMNLP 2021.
  13. (CCF B) Haolan Zhan, Lei Shen, Hongshen Chen and Hainan Zhang. CoLA: A Collaborative Latent Variable Model for Knowledge-Grounded Conversations.[C]//EMNLP 2021.
  14. (CCF B) Junpeng Liu, Yanyan Zou, Hainan Zhang, Hongshen Chen, Zhuoye Ding, Caixia Yuan and Xiaojie WANG. Topic-Aware Contrastive Learning for Abstractive Dialogue Summarization. [C]//EMNLP 2021.
  15. (CCF B) Haolan Zhan, Hainan Zhang#, Hongshen Chen, Lei Shen, Zhuoye Ding, Weipeng Yan, Yanyan Lan. Augmenting Knowledge-grounded Conversations with Sequential Knowledge Transition.[C]//NAACL 2021.
  16. (CCF A) Haolan Zhan, Hainan Zhang#, Hongshen Chen, Lei Shen, Yanyan Lan, Zhuoye Ding and Dawei Yin. Probing Product Description Generation via Posterior Distillation[C]//AAAI 2021.
  17. (CCF A) Haolan Zhan, Hainan Zhang#, Hongshen Chen, Lei Shen, Yanyan Lan, Zhuoye Ding and Dawei Yin. User-Inspired Posterior Network for Recommendation Reason Generation[C]//SIGIR 2020.
  18. (CCF A) Hainan Zhang, Yanyan Lan, Liang Pang, Hongshen Chen, Zhuoye Ding, Dawei Yin. Modeling Topical Relevance for Multi-Turn Dialogue Generation[C]//IJCAI 2020.
  19. (CCF A) Hainan Zhang, Yanyan Lan, Liang Pang, Jiafeng Guo, Xueqi Cheng. ReCoSa: Detecting the Relevant Contexts with Self-Attention for Multi-turn Dialogue Generation[C]//ACL 2019.
  20. (CCF A) Hainan Zhang, Yanyan Lan, Jiafeng Guo, Jun Xu, Xueqi Cheng. Reinforcing Coherence for Sequence to Sequence Model in Dialogue Generation[C]// IJCAI 2018.
  21. (CCF A) Hainan Zhang, Yanyan Lan, Jiafeng Guo, Jun Xu, Xueqi Cheng. Tailored sequence to sequence models to different conversation scenarios[C]//ACL 2018.
  22. Hainan Zhang, Yanyan Lan, Jiafeng Guo, Jun Xu, Xueqi Cheng. Neural or Statistical: An Empirical Study on Language Models for Chinese Input Recommendation on Mobile[J]//Lecture Notes of Computer Science. 2018.
  23. Hainan Zhang, Dayong Wu, Yue Liu, Xueqi Cheng. Chinese named entity recognition based on deep neural network[J].//Journal of Chinese Information Processing 31 (4), 28-35
  24. Hainan Zhang, Tonglei Guo, Yanyan Lan, Jiafeng Guo, Jianing Li and Xueqi Cheng. Gbot at the NTCIR-13 STC-2 Task.[C]//NTCIR 2017.
  25. 张海楠, 伍大勇, 刘悦, 程学旗. 基于深度神经网络的中文命名实体识别[J]. 中文信息学报, 2017, 31(4): 28-35.
  26. 郝长盈、兰艳艳、张海楠、郭嘉丰、徐君、程学旗. 基于拓展关键词信息的对话生成模型[J]. 山东大学学报（理工版）, 2019.

预印论文

  1. Qianchi Zhang, Hainan Zhang#, Liang Pang, Hongwei Zheng, Zhiming Zheng. AdaComp: Extractive Context Compression with Adaptive Predictor for Retrieval-Augmented Large Language Models[J]. arXiv preprint arXiv:2409.01579, 2024.
  2. Qianchi Zhang, Hainan Zhang#, Liang Pang, Hongwei Zheng, Yongxin Tong, Zhiming Zheng. FineFilter: A Fine-grained Noise Filtering Mechanism for Retrieval-Augmented Large Language Models[J]. arXiv preprint arXiv:2502.11811, 2025.

公开或授权专利 Patents

  1. 张海楠等. 一种基于强化学习的自适应聚合防御复杂投毒攻击方法. CN202411453984.9
  2. 张海楠等. 一种利用隐藏状态过滤的防隐私攻击方法及装置. CN202411512708.5
  3. 张海楠等. 一种区块链智能合约代码生成大型语言模型的方法. CN202411494298.6
  4. 张海楠等. 文案检测方法、装置、设备及计算机可读存储介质. CN202210952230.2
  5. 张海楠等. 模型训练方法、答复问题生成方法、装置、设备和介质. CN202310769116.0
  6. 张海楠等. 机器学习模型训练方法、摘要生成方法和装置. CN202210257581.1
  7. 张海楠等. 翻译模型的处理方法、装置和计算机可读存储介质. CN202210150760.5
  8. 张海楠等. 推荐理由生成方法、装置、设备及存储介质. CN202010619641.0 (已授权）
  9. 张海楠等. 用于确定信息的方法和装置. CN202210151466.6
  10. 张海楠等. 物品文案生成网络训练方法、物品文案生成方法、装置. CN202110084578.X
  11. 张海楠等. 文本输出方法、装置、设备以及存储介质. CN202110586299.3

在读学生 Current Students

已毕业学生 Graduated Students

与我联系

zhanghainan@buaa.edu.cn zhanghainan1990@163.com