当前位置: Home / Events / 2026
Haisheng Yang: Paradigm Reshaping: AI-Assisted Empirical Replication, Text Parsing, and Model Fine-Tuning in Economics
发布时间:2026-05-20 16:48:00 浏览次数:31

The 374th Wenlan Financial Forum

Topic

Paradigm Reshaping: AI-Assisted Empirical Replication, Text Parsing, and Model Fine-Tuning in Economics

Speaker:

Haisheng Yang, Professor

Lingnan College, Sun Yat-sen University

Host

Xianming Sun, Professor

School of Finance, Zhongnan University of Economics and Law

Innovation And Talent Base For Digital Technology And Finance

Time:

10:00-12:00, Friday, May 22, 2026

Location:

508 Conference Room, Wenquan South Building


Abstract:This lecture, themed "Economic Research in the Intelligent Era," provides a comprehensive overview of how artificial intelligence deeply empowers the entire workflow of fundamental economic research and empirical analysis. It will first explore how to leverage advanced AI programming environments (e.g., Claude Code) to efficiently process complex data and assist in writing Python and Stata scripts, thereby significantly shortening the replication cycle of classic empirical papers. Next, drawing on cutting-edge literature, the lecture will delve into how to use prompt engineering to guide large language models (LLMs) in accurately extracting micro‑dimensional structured data from massive unstructured text, while effectively addressing the hallucination problem of LLMs. In terms of literature review and framework construction, the lecture will demonstrate on‑site how to use proprietary knowledge base tools such as NotebookLM to quickly generate literature reviews and presentation slides based on a specified PDF library with zero fabrication, helping researchers overcome information overload. Finally, the lecture will focus on the local fine‑tuning of large models in the economics vertical domain, systematically introducing the LoRA technique and discussing how instruction fine‑tuning enables open‑source base models to deeply learn and capture specific economic laws, thereby creating exclusive measurement and prediction tools with economic "domain intuition." The lecture aims to help faculty and students break down technical barriers, seamlessly integrate cutting‑edge AI tools into daily research, and build a new type of research productivity.



Speaker Introduction

图片1.png

Haisheng Yang is a Professor and Doctoral Supervisor at Lingnan College, Sun Yat-sen University. His research interests include industrial policy evaluation, causal inference, AI and deep learning, and network analysis. He has published nearly 80 papers in leading academic journals, including over 30 papers in top-tier Chinese journals and SCI/SSCI journals such as Economic Research Journal (8 papers), Management World, China Economic Quarterly, Journal of Management Sciences in China, Expert Systems With Applications, ACM Computing Surveys, Ecological Economics, Accounting and Finance, Emerging Markets Review, Economic Modelling, and Journal of Financial Markets. Five of his papers have received awards: First Prize for Outstanding Research Paper in Philosophy and Social Sciences (9th Guangdong Province), Second Prize for Outstanding Financial Research Achievements (10th Guangdong Finance Society), and two Second Prizes for Outstanding Financial Research Achievements (12th Guangdong Province). His papers have accumulated over 5,000 citations on CNKI and Google Scholar, with more than 12,000 downloads. His work has sparked heated discussions at several major academic conferences, and one of his papers was accepted to IJCAI 2025.