Peng Yifeng bio photo

Email

My CV

LinkedIn

Github

About Me

It’s Peng here(or just call me Bob).

I am looking for 2026 Spring/Fall intake Ph.D in Finance/Fintech as well as artificial intellgence interdisciplinary works would be preferable.

Currently, I worked as a research asscioate under the guidence of Prof. CHEN chen in Chinese University of Hong Kong (Shenzhen), meanwhile I’m collaborating with researchers with similiar interests.

If you are interested in any aspect of me, I would love to chat and collaborate, email me at yifengpeng[at]link[dot]cuhk[dot]edu[dot]cn or schedule a Online Coffee Chat with me.

Academic Background

  • Sep 2021 - Jul 2023: Chinese University of Hong Kong (Shenzhen) (MSc, Finance)
  • Sep 2020 - Aug 2021: University of Hong Kong (MSc, Computer Science)
  • Sep 2016 - June 2020: Hunan University (BEng, Computer Science and Technology)

Research Interests

My current research focuses on the Alignment Problem in Large Language models (LLMs), I’m also a fan for Behavioral Finance, with a heart for bounded rationality. Hoping that AI can liberate productivity and benefit mankind.

Try to find a connection between human behavior and machine behavior, LLMs were trained by biased data (Researchs were done to make sure no harm to human, but how to avoid potential future harms when everyone use it everyday in multiply field?), and human were influenced by the biased information, we would face a future together with AI, how to make the AI more aligned with human values is the key point of my research.

News and Updates

  • December 23 2025: Our new preprint “CDiff: Bridging Conditional Diffusion and Causal Inference with Statistical Guarantees via Principled Distribution Approximation” is now available. We have submitted this work to ICML 2026. [Preprint]
  • December 10 2025: Thrilled to collaborate with Prof. Qianqian Xie’s team at WHU on our latest work “Hidden Memorization Trap in Financial LLMs.” This study explores the reliability of LLMs in financial contexts. [Draft]
  • November 10 2025: Achieved 17th place in the Kaggle Diabetes Prediction Challenge. Although I have limited time for extracurricular competitions this year, it was a great exercise!
  • August 18 2025: Our work “Poisoned Pills: insights from vulnerability disparity among llms” has been submitted to NeurIPS Lock-LLM Workshop 2025 for peer discussion.
  • March 1 2025: Happy to announce that our work “Poisoned Pills: insights from vulnerability disparity among llms” has been accepted by ACL 2025.
  • February 16 2025: Our work Poisoned Pills: insights from vulnerability disparity among llms, submitted to ACL 2025. [Preprint].
  • February 1 2025: I got 25/2722 in Kaggle Forcasting Sticker Sales Competition, really really really unbelievable works for me! Screenshot for result
  • January 1 2025: Happy New Year! An research about how fund distribution channel affect return submitted to 15th Financial Market and Corporate Governance Conference 2025 (FMCG 2025). Early Access
  • December 28 2024: I achieved new record in Kaggle Jane Street Real-Time Market Data Forecasting.
  • November 15 2024: An multi-disciplinary research in neuroeconomics with Gao Chen & Xi Gong was finished. Early Access(On working)
  • September 5 2024: I extended my internship at GEEKR as remote worker, need a balance between RA workload.
  • August 29 2024: Really thrilled be informed that My paper Sentiment towards Overtrading has been accepted by Journal of Finance and Accounting. Acceptance
  • July 4 2024: Finished the manuscript of Generative AI Goverance Framework to AI and Ethics, it’s a capstone for my Alignment Research.
  • June 26 2024: I finished my paper New intelligent empowerment for digital transformation🔗 and submiited to Technovation
  • May 21 2024: The Online Coffee Chat and disqus are available now, welcome to leave a message and chat with me!
  • April 21 2024: New Personl website! Welcome everyone!