AlphaRLM: Aligning Language Models for Literature-Grounded Symbolic Factor Discovery

Paper: arXiv 2610.14959

Authors: Jie Huang, Yuecheng Hong, Yao Chen, Hao Zhong, Linfeng Xiao, Xudong Liu, Hao Xia, Jianxuan Xin

Abstract

Formulaic alpha mining seeks to discover executable expressions that capture predictive market signals in a vast, noisy search space. Recent LLM-based approaches use prompting, iterative search, or seed-factor evolution, leaving direct factor construction through policy learning less explored. To address this, we introduce AlphaRLM, a reinforcement-learning framework that trains LLMs to construct executable symbolic factors from financial literature without input factor expressions. AlphaRLM incorporates Literature-Guided Operator Assembly (LEGO), which conditions factor construction on literature-derived ideas and executable operators, providing semantic priors for exploration. It further aligns the generation policy with market feedback through executable-validity and grouped-return-separation rewards that capture non-monotonic structures across factor quantiles. Extensive experiments on CSI 300, CSI 1000, and S&P 500 show that AlphaRLM outperforms strong factor-mining and LLM-based baselines in annualized excess return. Notably, the 8B model achieves strong cross-market performance, highlighting the promise of literature-conditioned, market-aligned factor generation.

Complexity vs Empirical Score

• Math Complexity: 7.5/10
• Empirical Rigor: 8.0/10
• Quadrant: Holy Grail — high math complexity, high empirical rigor

Why this score: This paper presents a novel reinforcement learning framework for symbolic factor discovery, integrating LLMs with market feedback. The empirical validation across multiple markets is strong, and the methodology is clearly articulated, though full reproducibility would depend on code/data availability not detailed in the excerpt.

Research Flowchart

  flowchart TD
    A[Research Goal: Discover Executable Expressions for Predictive Market Signals] --> B(Problem: Direct Factor Construction via Policy Learning Less Explored);
    B --> C{Methodology: AlphaRLM Framework};
    C --> D[Key Component 1: Literature-Guided Operator Assembly (LEGO)] --> E(Inputs: Financial Literature & Executable Operators);
    C --> F[Key Component 2: RL Alignment with Market Feedback] --> G(Rewards: Executable Validity & Grouped Return Separation);
    E & G --> H[Computational Process: Train LLMs for Factor Construction];
    H --> I[Outcomes: Outperforms Baselines in Annualized Excess Return];
    I --> J[Key Finding: 8B Model Achieves Strong Cross-Market Performance];