Topic: Deep Learning and Numerical Solution of Large-Scale Economic Models

Speaker: Assistant Professor Huang Ji, Department of Economics, The Chinese University of Hong Kong

Host: Professor Wu Ji from RIEM, SWUFE

Time: July 15, 2025 (Tuesday) 10:00-11:30

Location: Conference Room 1211, Gezhi Building, Liulin Campus, SWUFE

Organizer: RIEM

Speaker's Profile

        Huang Ji is currently an Assistant Professor in the Department of Economics at the Chinese University of Hong Kong. He received his Bachelor of Management from Southwestern University of Finance and Economics in 2006, Master of Economics degree from Nankai University in 2009, and his Ph.D. in Economics from Princeton University in 2015. From July 2015 to July 2018,he was on the faculty of the Department of Economics at the National University of Singapore. His earlier research covered shadow banking and macro‑finance, with work on shadow banking published in journals such as Journal of Economic Theory and Review of Finance. His current research focuses on numerical solution methods for high‑dimensional, continuous‑time equilibrium models based on machine learning and backward stochastic differential equations, with a long‑term aim of commercializing large‑scale economic models.

Abstract

 Although numerical computation has broadened the scope of economic research, but many economic models remain simplified by unrealistic assumptions to avoid the “curse of dimensionality” and keep models solvable and parameters estimable. Advances in computing hardware and software, especially deep learning and artificial intelligence, have shown transformative, industrial‑scale capabilities in overcoming dimensionality constraints, as seen in AlphaGo and ChatGPT. This lecture aims to introduce the key technical principles of deep learning (including deep reinforcement learning) and discuss their applications to the numerical solution of large‑scale economic models.