ChengXing Xie

ChengXing Xie

Artificial Intelligence Student

Xidian University

Camel.ai

Biography

I am a fourth-year Computer Science student at Xidian University, with visiting research experiences at HKU and KAUST. Currently, I am interning at Shanghai AI Lab, focusing on improving open-source LLMs for real-world tasks.

My research interests lie in LLM reasoning and real-world task solving using LLMs. I am particularly fascinated by how LLMs can be enhanced to perform complex reasoning and planning in practical applications. I believe there is great potential in refining LLMs to bridge the gap between language models and consistent decision-making, and I aim to explore this further in my future research.

For further information or collaboration opportunities, feel free to reach out to me at xiechengxing34@gmail.com.

Interests
  • LLM Reasoning
  • Real-World Task Solving by LLM
  • Generative Model
Education
  • B.E in Computer Science, 2021-now

    Xidian University

  • Visiting Student, 2023.7-2024.1

    KAUST (King Abdullah University of Science and Technology)

Recent Publications

Quickly discover relevant content by filtering publications.
(2024). Can Large Language Model Agents Simulate Human Trust Behaviors?. Accepted at NeurIPS 2024.

PDF Code

(2023). TFRGAN Leveraging Text Information for Blind Face Restoration with Extreme Degradation.

PDF

Experience

 
 
 
 
 
Shanghai AI Lab
Research Intern at Shanghai AI Lab
July 2024 – Present Shanghai, China
Focused on enhancing the performance of open-source LLMs in real-world code-related tasks.
 
 
 
 
 
The University of Hong Kong (HKU)
Research Assistant at HKU
February 2024 – May 2024 Hong Kong, China
Working under the supervision of Prof. Difan Zou on multi-agent planning and improving LLM agents’ problem-solving abilities.
 
 
 
 
 
Image and Video Understanding Lab
Visiting Student in KAUST (IVUL, Advisor PhD Guohao Li, Prof. Bernard Ghanem)
July 2023 – January 2024 Thuwal, Saudi Arabia
Working on the Camel project, probing agent trust behavior. The paper titled “Can Large Language Models Simulate Human Trust Behaviors?” has been accepted in NeurIPS 2024.
 
 
 
 
 
Sensetime
Research Intern
February 2022 – November 2022 Xian China
 
 
 
 
 
Xidian University
Undergraduate of Xidian University
September 2021 – June 2025 Xian China
Computer Science Undergraduate Student

Projects

*
TFRGAN Leveraging Text Information for Blind Face Restoration with Extreme Degradation
Using text information to achieve better res in BFR task.(This work has been submitted to CVPR workshop)
Sleep State Analysis
Using AI to monitor human’s sleep states.(SenseTime Intern)
Cough Detection
Using CNN to do binary classification for elder’s sounds. (SenseTime Intern)
Audio Clip
Using Audio and Text do Clip training.(SenseTime Intern)
Animation Picture Generation
Using a deep convolutional generative adversarial network (DCGAN) generate anime-style avatar images.

Contact

If you want to contact me, free free to contact me at any time