Federico Fontana, Chief Technology Officer at XAI
15 Sept 2023
At XAI Asset Management, we're constantly pushing the boundaries of technology to achieve excellence in finance. Today, we're thrilled to introduce you to our event-driven market simulator used internally to train reinforcement learning agents. A component of our technological infrastructure, now open source on GitHub .
Reinforcement Learning (RL) is one of three basic machine learning paradigms, alongside supervised learning and unsupervised learning. RL finds applications in finance due to its ability to optimise sequential decision-making in dynamic environments. By training agents to learn from historical or synthetic data and adapt to changing market conditions, RL enables the development of trading algorithms that can identify patterns and manage risks.
The RL landscape is vast. At the core, applications require a tasks to solve (also known as environments) and RL algorithms. In order to facilitate progress, the research community has adopted OpenAI/gym  as the standard protocol to design, implement and share environments. Therefore, the implementation of many RL algorithms can be used off the shelf with custom environments built upon OpenAI/gym.
Our contribution consists in simplifying the journey for RL researchers in building financial applications. By providing this market simulator, we reduce the time and effort spent on building infrastructure, allowing users to spend more time on the actual project. This means users can leverage and scale open-sourced RL algorithms, without the need to start from scratch.
 Brockman, G., Cheung, V., Pettersson, L., Schneider, J., Schulman, J., Tang, J. and Zaremba, W., 2016. Openai gym. arXiv preprint arXiv:1606.01540.