FinSMART Optimizes Financial Sentiment via Market-Aligned Reinforcement Learning

FinSMART is a reinforcement learning framework that optimizes financial sentiment signals using realized market outcomes rather than static, supervised datasets. It employs a signal extraction pipeline with discrete asymmetric trading rewards to handle the non-stationary nature of financial markets.