Trade-R1: Bridging Verifiable Rewards to Stochastic Environments via Process-Level Reasoning Verification
This work addresses the problem of reward hacking in RL for financial domains, offering a domain-specific solution that is incremental by building on existing verification and reward integration methods.
The paper tackles the challenge of applying reinforcement learning to financial decision-making in stochastic markets, where noisy rewards cause reward hacking, by proposing Trade-R1, a framework that uses process-level reasoning verification to bridge verifiable rewards, resulting in reduced reward hacking and superior cross-market generalization with high reasoning consistency.
Reinforcement Learning (RL) has enabled Large Language Models (LLMs) to achieve remarkable reasoning in domains like mathematics and coding, where verifiable rewards provide clear signals. However, extending this paradigm to financial decision is challenged by the market's stochastic nature: rewards are verifiable but inherently noisy, causing standard RL to degenerate into reward hacking. To address this, we propose Trade-R1, a model training framework that bridges verifiable rewards to stochastic environments via process-level reasoning verification. Our key innovation is a verification method that transforms the problem of evaluating reasoning over lengthy financial documents into a structured Retrieval-Augmented Generation (RAG) task. We construct a triangular consistency metric, assessing pairwise alignment between retrieved evidence, reasoning chains, and decisions to serve as a validity filter for noisy market returns. We explore two reward integration strategies: Fixed-effect Semantic Reward (FSR) for stable alignment signals, and Dynamic-effect Semantic Reward (DSR) for coupled magnitude optimization. Experiments on different country asset selection demonstrate that our paradigm reduces reward hacking, with DSR achieving superior cross-market generalization while maintaining the highest reasoning consistency.