AlphaPROBE: Alpha Mining via Principled Retrieval and On-graph biased evolution

Extracting signals through alpha factor mining is a fundamental challenge in quantitative finance. Existing automated methods primarily follow two paradigms: Decoupled Factor Generation, which treats factor discovery as isolated events, and Iterative Factor Evolution, which focuses on local parent-child refinements. However, both paradigms lack a global structural view, often treating factor pools as unstructured collections or fragmented chains, which leads to redundant search and limited diversity. To address these limitations, we introduce AlphaPROBE (Alpha Mining via Principled Retrieval and On-graph Biased Evolution), a framework that reframes alpha mining as the strategic navigation of a Directed Acyclic Graph (DAG). By modeling factors as nodes and evolutionary links as edges, AlphaPROBE treats the factor pool as a dynamic, interconnected ecosystem. The framework consists of two core components: a Bayesian Factor Retriever that identifies high-potential seeds by balancing exploitation and exploration through a posterior probability model, and a DAG-aware Factor Generator that leverages the full ancestral trace of factors to produce context-aware, nonredundant optimizations. Extensive experiments on three major Chinese stock market datasets against 8 competitive baselines demonstrate that AlphaPROBE significantly gains enhanced performance in predictive accuracy, return stability and training efficiency. Our results confirm that leveraging global evolutionary topology is essential for efficient and robust automated alpha discovery. We have open-sourced our implementation at https://github.com/gta0804/AlphaPROBE.

💡 Research Summary

AlphaPROBE tackles the long‑standing problem of automated alpha factor discovery by reframing it as a strategic navigation problem on a Directed Acyclic Graph (DAG). In this graph, each node represents an alpha factor (a mathematical expression that maps raw market data to a predictive signal) and each directed edge encodes an evolutionary relationship – a child factor derived from a parent. This global structural view contrasts sharply with the two dominant paradigms in the literature: Decoupled Factor Generation (DFG), which treats factor creation as independent sampling events, and Iterative Factor Evolution (IFE), which refines factors locally along parent‑child chains. Both paradigms ignore the broader topology of the factor pool, leading to redundant searches, limited diversity, and inefficient use of previously discovered knowledge.

The AlphaPROBE system consists of two tightly coupled components:

1. Bayesian Factor Retriever – This module selects the most promising parent factors for the next generation. It formulates the selection as a posterior probability maximization:
   \