AlphaPortfolio: Direct Construction Through Deep Reinforcement Learning and Interpretable AI

76 Pages Posted: 20 Apr 2020 Last revised: 2 Mar 2022

See all articles by Lin William Cong

Lin William Cong

Cornell University - Samuel Curtis Johnson Graduate School of Management; National Bureau of Economic Research (NBER)

Ke Tang

Institute of Economics, School of Social Sciences, Tsinghua University

Jingyuan Wang

Beihang University (BUAA)

Yang Zhang

Beihang University (BUAA)

Date Written: August 1, 2021

Abstract

We directly optimize the objectives of portfolio management via deep reinforcement learning---an alternative to conventional supervised-learning paradigms that routinely entail first-step estimations of return distributions or risk premia. We develop multi-sequence, attention-based neural-network models tailored for the distinguishing features of financial big data, while allowing interactions with the market states and training without labels. Such AlphaPortfolio models yield stellar out-of-sample performances (e.g., Sharpe ratio above two and over 13% risk-adjusted alpha with monthly re-balancing) that are robust under various market conditions and economic restrictions (e.g., exclusion of small and illiquid stocks). We further demonstrate AlphaPortfolio's flexibility to incorporate transaction costs, state interactions, and alternative objectives, before applying polynomial-feature-sensitivity analysis to uncover key drivers of investment performance, including their rotation and nonlinearity. Overall, we highlight the utility of deep reinforcement learning in finance and "economic distillation" for model interpretation.

Keywords: Artificial Intelligence, Distillation, LSTM, Machine Learning, Portfolio Theory, Reinforcement Learning.

Suggested Citation

Cong, Lin and Tang, Ke and Wang, Jingyuan and Zhang, Yang, AlphaPortfolio: Direct Construction Through Deep Reinforcement Learning and Interpretable AI (August 1, 2021). Available at SSRN: https://ssrn.com/abstract=3554486 or http://dx.doi.org/10.2139/ssrn.3554486

Lin Cong (Contact Author)

Cornell University - Samuel Curtis Johnson Graduate School of Management ( email )

Ithaca, NY 14853
United States

HOME PAGE: http://www.linwilliamcong.com/

National Bureau of Economic Research (NBER) ( email )

1050 Massachusetts Avenue
Cambridge, MA 02138
United States

Ke Tang

Institute of Economics, School of Social Sciences, Tsinghua University ( email )

No.1 Tsinghua Garden
Beijing, 100084
China

Jingyuan Wang

Beihang University (BUAA) ( email )

37 Xue Yuan Road
Beijing 100083
China

Yang Zhang

Beihang University (BUAA) ( email )

37 Xue Yuan Road
Beijing 100083
China

Do you have a job opening that you would like to promote on SSRN?

Paper statistics

Downloads
3,730
Abstract Views
10,526
rank
4,064
PlumX Metrics