FinFlowRL: An Imitation-Reinforcement Learning Framework for Adaptive Stochastic Control in Finance
This addresses the challenge of non-stationary financial markets for quantitative finance practitioners, though it appears incremental as it builds on existing imitation and reinforcement learning techniques.
The paper tackles the problem of traditional stochastic control methods performing poorly in real-world financial markets due to their reliance on simplifying assumptions, by introducing FinFlowRL, a framework that combines imitation and reinforcement learning to adaptively optimize control policies, consistently outperforming expert strategies across diverse market conditions.
Traditional stochastic control methods in finance struggle in real world markets due to their reliance on simplifying assumptions and stylized frameworks. Such methods typically perform well in specific, well defined environments but yield suboptimal results in changed, non stationary ones. We introduce FinFlowRL, a novel framework for financial optimal stochastic control. The framework pretrains an adaptive meta policy learning from multiple expert strategies, then finetunes through reinforcement learning in the noise space to optimize the generative process. By employing action chunking generating action sequences rather than single decisions, it addresses the non Markovian nature of markets. FinFlowRL consistently outperforms individually optimized experts across diverse market conditions.