Deep Reinforcement Learning for Active Trading: LLaMA 3.2 1B Powers Trading Decisions

Researchers have developed a novel approach for active trading using deep reinforcement learning, specifically for Bitcoin and Tesla assets. The system employs four distinct deep reinforcement learning algorithms: Policy Gradient, Proximal Policy Optimization, Deep Q-Learning, and Deep Deterministic Policy Gradient. These agents leverage technical indicators, cyclical calendar encodings, and sentiment scores derived from LLaMA 3.2 1B to make daily trading decisions. To enhance generalization and performance, an alpha reward mechanism was introduced, and hyperparameters were optimized through extensive trials. The Deep Deterministic Policy Gradient algorithm demonstrated superior performance on the test set, while Deep Q-Learning was pre-selected for its high validation Sharpe ratio. AI

IMPACT This research demonstrates a sophisticated application of LLMs and reinforcement learning for financial market prediction, potentially influencing algorithmic trading strategies.

RANK_REASON The cluster contains a research paper detailing a novel application of deep reinforcement learning for financial trading. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.LG →

• Bitcoin
• CLEF 2026 FinMMEval Lab
• Deep Deterministic Policy Gradient
• Deep Q-Learning
• LLaMA 3.2 1B
• Policy Gradient
• Proximal Policy Optimization
• Tesla

• paper
• product

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →