FLAG-Trader: Fusion LLM-Agent with Gradient-based Reinforcement Learning for Financial Trading

TL;DR

FLAG-Trader combines large language models with gradient-based reinforcement learning to improve decision-making in financial trading, leveraging pre-trained knowledge and domain adaptation for better multi-step, goal-oriented trading strategies.

Contribution

The paper introduces a novel unified architecture that integrates LLMs with reinforcement learning for financial trading, enabling improved multi-step decision-making and domain adaptation.

Findings

Enhanced trading performance through RL optimization.

Improved results on financial-domain tasks.

Effective domain adaptation with parameter-efficient fine-tuning.

Abstract

Large language models (LLMs) fine-tuned on multimodal financial data have demonstrated impressive reasoning capabilities in various financial tasks. However, they often struggle with multi-step, goal-oriented scenarios in interactive financial markets, such as trading, where complex agentic approaches are required to improve decision-making. To address this, we propose \textsc{FLAG-Trader}, a unified architecture integrating linguistic processing (via LLMs) with gradient-driven reinforcement learning (RL) policy optimization, in which a partially fine-tuned LLM acts as the policy network, leveraging pre-trained knowledge while adapting to the financial domain through parameter-efficient fine-tuning. Through policy gradient optimization driven by trading rewards, our framework not only enhances LLM performance in trading but also improves results on other financial-domain tasks. We present extensive empirical evidence to validate these enhancements.