Cheng Qian

Cheng Qian

Undergrad@Tsinghua, THUNLP qianc20@mails.tsinghua.edu.cn

Cheng Qian

This is Cheng Qian’s personal homepage.

A short introduction

I am currently a Ph.D. student at University of Illinois at Urbana-Champaign, advised by Heng Ji. Previously, I am undergraduate student studying at Tsinghua University, majoring in Computer Science and Technology and advised by Zhiyuan Liu.

My research interests primarily lie in the field of natural language processing, with a particular focus on tool learning and tool creation of large language models (LLMs), LLM-driven AI agents and its application in science, elicition and interpretability of LLM knowledge.

In 2023 summer, I worked with Sherry Wu on the impacts of external knowledge (e.g. from tools, retrieval, etc.) to parametric knowledge (paper). In previous research, I have investigated the tool creation ability of large language models (paper), and how the tool-use ability could be adapted to open-source models (paper). This 2024 spring, I have also investigated into how dialogue agent could understand human’s implicit intention (paper).

Before 2023, I have explored the concept of recycling outdated weights during the lifelong knowledge acquisition of LLMs, which was published in ACL 2023 findings (paper). Additionally, I have contributed to research on the mode connectivity of models’ various minima, which was presented at EMNLP 2022 (paper).

I am starting my Ph.D. from Fall 2024. If you are interested in my research or would like to collaborate, please feel free to reach out to me! (I am also actively seeking for internship opportunities in 2025 summer!)

Recent Research Focuses:

  • Build LLM to understand human’s implicit intention in dialogue.
  • Propose CREATOR to help automatic tool-creation.
  • Adapt tool-using ability to open-source LLaMA models.
  • Investigate the impact of external knowledge (e.g. from tools, retrieval, etc.) brought to LLMs.

Latest News

  • 2024.9: Starting a new PhD life at UIUC! Excited to work with Prof. Heng Ji!
  • 2024.8: I am currently @ ACL 2024 to present the paper Tell Me More here! Greetings to everyone~
  • 2024.6: New paper Toolink presented at NAACL!
  • 2023.10: The code for CREATOR is publicly released! Check from the repository here!
  • 2023.9: The latest paper about knowledge conflic is release on ArXiv here! Many thanks to Sherry Wu and Xinran Zhao. It has been a wonderful summer at CMU!
  • 2023.7: @ ACL 2023 in Toronto! My first time participate a conference offline!

Papers

(*indicates equal contribution)

Cheng Qian, Bingxiang He, Zhong Zhuang, Jia Deng, Yujia Qin, Xin Cong, Zhong Zhang, Jie Zhou, Yankai Lin, Zhiyuan Liu, Maosong Sun. ACL 2024 Main. (Paper)

Cheng Qian, Xinran Zhao, Sherry Tongshuang Wu. “Merge Conflicts!” Exploring the Impacts of External Distractors to Parametric Knowledge Graphs. COLM 2024. (Paper / Code)

Cheng Qian, Chi Han, Yi R. Fung, Yujia Qin, Zhiyuan Liu, Heng Ji. Tool Creation for Disentangling Abstract and Concrete Reasonings of Large Language Models. EMNLP 2023 Findings. (Paper / Code)

Cheng Qian, Chenyan Xiong, Zhenghao Liu, Zhiyuan Liu. Toolink: Linking Toolkit Creation and Using through Chain-of-Solving on Open-Source Model. NAACL 2024 Main. (Paper / Code)

Yujia Qin, Shengding Hu, Yankai Lin, Weize Chen, Ning Ding, Ganqu Cui, Zheni Zeng, Yufei Huang, Chaojun Xiao, Chi Han, Yi Ren Fung, Yusheng Su, Huadong Wang, Cheng Qian, Runchu Tian, Kunlun Zhu, Shihao Liang, Xingyu Shen, Bokai Xu, Zhen Zhang, Yining Ye, Bowen Li, Ziwei Tang, Jing Yi, Yuzhang Zhu, Zhenning Dai, Lan Yan, Xin Cong, Yaxi Lu, Weilin Zhao, Yuxiang Huang, Junxi Yan, Xu Han, Xian Sun, Dahai Li, Jason Phang, Cheng Yang, Tongshuang Wu, Heng Ji, Zhiyuan Liu, Maosong Sun. Tool Learning with Foundation Models. ACM COMPUTING SURVEYS.(Paper / Code)

Cheng Qian*, Yujia Qin*, Yankai Lin, Xu Han, Zhiyuan Liu, Maosong Sun and Jie Zhou. Recyclable Tuning for Continual Pre-training. ACL 2023 findings. (Paper / Code)

Cheng Qian*, Yujia Qin*, Jing Yi*, Weize Chen, Yankai Lin, Xu Han, Zhiyuan Liu, Maosong Sun and Jie Zhou. Exploring Mode Connectivity for Pre-trained Language Models. EMNLP 2022. (Paper / Code)

For more information

More info about Cheng Qian can be found here! (downloaded CV).