Can Reinforcement Learning Efficiently Discover Price Manipulation?
Updated
Updated · arxiv.org · Jul 7
Can Reinforcement Learning Efficiently Discover Price Manipulation?
1 articles · Updated · arxiv.org · Jul 7
Summary
A new study finds that reinforcement learning (RL) can efficiently uncover price manipulation strategies in financial markets with non-linear price impact.
Researchers compared RL agents to model-based approaches and found RL often outperformed traditional methods, especially when parameter estimates were noisy or data was limited.
These findings highlight both the power and risks of RL in finance, underscoring the need for regulatory safeguards to prevent unintended market manipulation.