Evolution of machine learning in financial risk management: A survey

Financial risk management plays a crucial role in daily financial decision-making, aiming to mitigate risk and maximize profit. Given its reliance on data, financial risk management can greatly benefit from the application of machine learning tools. Over the years, we've observed a clear trend in the evolution of these applications, marked by increasing model complexity and a broader range of manageable tasks. This paper contributes to the field in three key dimensions: First, we provide a clear taxonomy of risks and an introduction to relevant machine learning methods to establish a foundation and identify the targeted issues. Next, we explore real-world data applications, discussing the pros and cons of three methods, from the earliest to the most recent. Finally, based on the observed results, we highlight current challenges and limitations and propose potential directions for improvement.