Humans don't learn from millions of labeled examples. Instead, we often learn from positive or negative experiences that we associate with our actions. Children that touch a hot stove once will never touch it again. Learning from experiences and the associated rewards or punishments is the core idea behind reinforcement learning (RL). RL allows us to learn sophisticated decision-making rules while having no data at all. Through this approach, several high-profile breakthroughs occurred in AI, such as AlphaGo, which beat the world Go champion in 2016.
In finance, reinforcement learning, also known as RL, is making inroads as well. In its 2017 report, Machine learning in investment management (https://www.ahl.com/machine-learning), Man AHL outlined a reinforcement system for order routing in the FX and futures market. Order routing is a classic problem in quantitative finance. When placing an order, funds can usually choose from different...