Organizations: Beijing University of Posts and Telecommunications, China · Beijing Value Simplex Technology Co., Ltd., China · Yangtze Delta Research Institute, University of Electronic Science and Technology of China, China · Wuhan Eonray Technology Co., Ltd., China
We study alpha factor mining, the automated discovery of predictive signals from noisy, non-stationary market data-under a practical requirement that mined factors be directly executable and auditable, and that the discovery process remain computationally tractable at scale. Existing symbolic approaches are limited by bounded expressiveness, while neural forecasters often trade interpretability for performance and remain vulnerable to regime shifts and overfitting. We introduce FactorEngine (FE), a program-level factor discovery framework that casts factors as Turing-complete code and improves both effectiveness and efficiency via three separations: (i) logic revision vs. parameter optimization, (ii) LLM-guided directional search vs. Bayesian hyperparameter search, and (iii) LLM usage vs. local computation. FE further incorporates a knowledge-infused bootstrapping module that transforms unstructured financial reports into executable factor programs through a closed-loop multi-agent extraction-verification-code-generation pipeline, and an experience knowledge base that supports trajectory-aware refinement (including learning from failures). Across extensive backtests on real-world OHLCV data, FE produces factors with substantially stronger predictive stability and portfolio impact-for example, higher IC/ICIR (and Rank IC/ICIR) and improved AR/Sharpe, than baseline methods, achieving state-of-the-art predictive and portfolio performance.
Figures & tables
Fig. 1: Overview of FactorEngine (FE). Left: Bootstrapping extracts factor ideas and converts pseudocode into executable Python to seed a knowledge-infused pool. Center: Evolution performs macro–micro co-evolution: LLM agents propose macro mutations guided by chains of experience, and Bayesian search conducts micro-level instantiation with fast validation and feedback updates. Right: Integration selects elite factors to train models for backtesting, producing portfolio-level feedback.
Fig. 2: Overview of the Bootstrapping module with a closed-loop extraction–verification–generation pipeline.
Fig. 3: (Left/Middle) Net cumulative excess returns in CSI300 and CSI500 under the cost-aware trading protocol. (Right) Factor-pool diversity diagnostic on CSI500 via MDS on 1−∣ρ∣ , with RoG and keep ratio reported in the legend.
Fig. 4: Yearly IC (Left) and Rank IC (Middle) comparisons on the CSI300 market. Mean IC and Rank IC between the top 10% factors and future returns at T+N on the CSI300 market across three settings (Right; the plot label “T+N” denotes an N-day-ahead return horizon).
Fig. 5: Evolution stability of FE-alpha on CSI300. Left: mean test-set ∣fitness∣ trajectory across 20 initial seed programs, with shading indicating across-seed dispersion. Right: generation-level test-set ∣fitness∣ distribution across observed nodes.
Fig. 6: Left : Effect of Bayesian micro-instantiation. bay_avg uses local Bayesian instantiation, while no_bay_avg uses fixed/default parameters. Right : Backbone robustness across GPT-4o and Gemini-2.5-flash-lite.
config
RIC
RICIR
AR
IR
|MDD|
5alpha,1island,CoE
0.0325
0.2165
0.0728
0.6737
0.1678
5alpha,1island,top-k
0.0319
0.2125
0.0696
0.6428
0.1626
5alpha,2island,CoE
0.0346
0.2308
0.0888
0.7886
0.1691
5alpha,2island,top-k
0.0332
0.2189
0.0775
0.7085
0.1614
10alpha,1island,CoE
0.0344
0.2358
0.0782
0.7079
0.1787
10alpha,1island,top-k
0.0341
0.2266
0.0761
0.6944
0.1673
TABLE III: Comparisons of FE-alpha under different prompt settings, island configurations, and numbers of initial factors.
Fig. 7: Abandoned-Baby evolution quality. Bubbles show mean training ∣fitness∣ by generation, with radius encoding within-generation standard deviation. The inset reports valid Bayesian-instantiation scores for the selected generation-4 node.
Factor mining seeks to discover signals from financial data that predict future asset returns and guide portfolio construction. Automated factor mining now spans genetic programming, reinforcement learning, generative models, and large language model agents. Yet it remains unclear whether advances across these paradigms yield more generalizable, distinct, and economically useful financial signals. We introduce FactorBench, a portfolio-aware benchmark comparing roughly five thousand mined factors from nine automated mining methods across five equity markets. A shared data and evaluation contract supports both symbolic expressions and executable Python factors, connecting heterogeneous discovery algorithms to common signal combination and portfolio construction procedures. FactorBench traces the outputs of mining systems across three levels: factor validity, temporal generalization, and predictiveness beyond measured risk and style exposures; within- and across-method pool distinctness, including similarity to the benchmark Alpha101; and composite-signal quality and after-cost long-only and long--short portfolio performance. After systematically assessing whether advances in factor mining translate into signal quality and portfolio performance, FactorBench finds that no paradigm consistently dominates.
Zhuohan Wang, Carmine Ventre
Department of Informatics King’s College London London, UK
Formulaic alpha factor mining is a critical yet challenging task in quantitative investment, characterized by a vast search space and the need for domain-informed, interpretable signals. However, finding novel signals becomes increasingly difficult as the library grows due to high redundancy. We propose FactorMiner, a lightweight and flexible self-evolving agent framework designed to navigate this complex landscape through continuous knowledge accumulation. FactorMiner combines a Modular Skill Architecture that encapsulates systematic financial evaluation into executable tools with a structured Experience Memory that distills historical mining trials into actionable insights (successful patterns and failure constraints). By instantiating the Ralph Loop paradigm -- retrieve, generate, evaluate, and distill -- FactorMiner iteratively uses memory priors to guide exploration, reducing redundant search while focusing on promising directions. Experiments on multiple datasets across different assets and Markets show that FactorMiner constructs a diverse library of high-quality factors with competitive performance, while maintaining low redundancy among factors as the library scales. Overall, FactorMiner provides a practical approach to scalable discovery of interpretable formulaic alpha factors under the "Correlation Red Sea" constraint.
Yanlong Wang, Jian Xu, Hongkang Zhang +3
Tsinghua Shenzhen International Graduate School, Tsinghua University · Pengcheng Laboratory · Shenzhen Finance Institute, Chinese University of Hong Kong
Factor investing is ultimately grounded in market logic - the latent mechanism behind observed alpha factors that explains why they should persist across assets and regimes. However, recent factor mining prioritizes factor discovery over logic discovery, producing complex alpha factors with unclear rationale, while market logic remains largely handcrafted and difficult to scale. To address this challenge, we propose AlphaLogics, a market logic-driven multi-agent system for factor mining. AlphaLogics consists of three key components: (i) Market Logic Mining: reverse-extracting market logic from historical factor libraries to construct an initial market logic library; (ii) Factor Generation and Optimization: using new market logics generated in (i) to guide factor generation, and optimizing factors with backtesting feedback; and (iii) Market Logic Generation and Optimization: generating new market logics conditioned on the initial market logic library, and refining each market logic by aggregating the backtest outcomes of its guided factors, continuously refreshing the library. Experiments on CSI 500 and S&P 500 show that AlphaLogics consistently improves predictive metrics and risk-adjusted returns over representative baselines, while producing a market logic library that remains empirically useful for guiding further factor discovery.
Zhangyuhua Weng, Shengli Zhang, Taotao Wang +1
College of Electronic and Information Engineering, Shenzhen University