Reinforcement Learning in Finance

“The major winners will be financial services companies that embrace technology." ~ Alexander Peh, PayPal and Braintree

Over the last few years, Reinforcement Learning (RL) has gained significant attention as a framework for learning optimal decisions even in complex environments. Moreover, RL is currently considered one of the most promising research areas in machine learning and has already demonstrated immense potential for solving sophisticated real-world problems, such as in robotics, resource management or health care.

Recently, this methodology was introduced successfully at JP Morgan within their multi-billion USD derivatives trading business. It is expected that RL algorithms have the potential to disrupt the financial service industry significantly and revolutionize the development of new financial market products.

However, despite all major academic breakthroughs a broader deployment of RL in financial applications is currently lacking, and practical implementations mostly remain hidden as intellectual property of only the largest global players. One fundamental reason for this are the significant hurdles that practitioners are still facing in building such models in real-world settings.

Our goal is to overcome these hurdles by identifying and addressing outstanding practical challenges when applying RL and building exemplary and openly accessible use cases that showcase the vast potential of RL applications in finance.

These implementations will act as a starting point to further develop context-specific applications and new financial products. We specifically aim at making this exciting field available to SMEs in Switzerland thereby facilitating a significant cost reduction for businesses who want to employ RL solutions within their current business models. In order not to lose an edge on the highly competitive global stage, Swiss financial SMEs will have to be granted easier access to disruptive RL applications - this is what we propose in this project.  

More information about the project can be found at https://www.zhaw.ch/en/research/research-database/project-detailview/projektid/4582