Zoe Wang, who received her Master of Engineering degree in financial engineering in December 2020, published a paper in the summer 2021 issue of The Journal of Financial Data Science, arguably the top quant publication.
Wang’s paper, titled “Deep Q-Learning for Trading Cryptocurrency” was co-authored with Alex Fleiss, the CEO of Rebellion Research in New York City and Yu-Chien (Calvin) Ma, a Rutgers Business School graduate.
“As a result of this article being published, our hope is to get more fellow Cornell students to become interested in the digital currency/quantitative finance world,” said Wang, who interned at Rebellion Research during the summer of 2020. “This article could also provide some great insights to them, as cryptocurrencies are getting increasing attention from the big institutions worldwide.”
“The potential for applying various forms of both deep learning and reinforcement learning to trading cryptocurrencies is very exciting,” Fleiss said. “We have demonstrated an ability to create outsized returns and alpha in the alternative currency space and our research should be a building block for future work in the space. As always I am grateful for the support of Cornell Financial Engineering and (Cornell Financial Engineering Manhattan Director) Victoria Averbukh.”
The article sets forth a framework for deep reinforcement learning as applied to trading cryptocurrencies. Specifically, the authors adopt Q-Learning, which is a model-free reinforcement learning algorithm, to implement a deep neural network to approximate the best possible states and actions to take in the cryptocurrency market. Bitcoin, Ethereum, and Litecoin were selected as representatives to test the model. The Deep Q trading agent generated an average portfolio return of 65.98%, although it showed extreme volatility over the 2,000 runs. Despite the high volatility of deep reinforcement learning, the experiment demonstrates that it has exceptionally high potential to be employed and provides a solid foundation on which to build further research.
The key findings of the study were:
The authors use deep neural networks to create a Deep Q-Learning trading agent that approximates the best actions to take based on rewards to maximize returns from trading the three cryptocurrencies with the largest market capitalization.
The Deep Q-Learning agent generates a return of nearly 66% on average over the course of 2,000 episodes; however, the returns do exhibit a large standard deviation given the highly volatile nature of the cryptocurrencies.
The authors introduce a framework on which future deep reinforcement learning and rewards-based trading agents can be built and improved.
Following her graduation from Cornell, Wang took a position as an analyst with Atlas Capital Team L.P. in New York City.