Reproduce “AlphaPADI: Formulaic Alpha Discovery via Pool-Aware Hierarchical Discrete Diffusion” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Yanzheng Jin, Pengyang Shao, Yunshan Ma, Haowen Pan, Naixin Zhai, Chen-Hui Song, Fei Shen, Kenji Kawaguchi
AlphaPADI searches for stock-ranking formulas by reconstructing masked parts of a pool using its other formulas as context, then applying pool-level predictive and redundancy feedback to refine candidates and train the generator. On the CSI300 test set, it achieves a RIC of 0.0659, versus 0.0544 for AlphaQCM, a 21.1% improvement; the historical backtest does not establish future returns or live-trading performance.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.