AIQF Daily digest — 11 new papers

Quant Paper Radar · 3 of 11

1. 1

   Conditional value-at-risk under reward-penalty mechanism with applications to robust portfolio management

   Jun Cai, Tiantian Mao, Zhiqiao Song · ARXIV · Robust portfolio optimization

   we backtested it — Sharpe 0.97 · quant relevance 0.98

   The paper proposes a distributionally robust portfolio-allocation method that minimizes worst-case CVaR while accounting for returns and losses beyond a threshold. In experiments using real market data, its portfolios outperform several comparator models.

   original paper

2. 2

   Residual Learning in Empirical Asset Pricing

   Dexin Peng, Xiaoyu Wang · ARXIV · Machine-learning cross-sectional equity

   quant relevance 0.98

   The paper tests whether residual neural networks improve one-month-ahead US stock return forecasts and value-weighted long-short portfolio performance. It reports an out-of-sample Sharpe ratio of 2.07 for deep residual models, compared with 1.92 for shallow models. A backtest here could test that architecture comparison, but its shorter history and available predictor set may not reproduce those figures.

   original paper

3. 3

   STOCK-JEPA: Prior-Anchored Latent Revision Representation Learning in Equity Markets

   Yizhi Luo, Jiahe Yi, Jianhui Zhang et al. · ARXIV · Machine-learning equity ranking

   quant relevance 0.96

   Stock-JEPA learns equity representations for return ranking and evaluates their long-short portfolio performance. Its US equity application can be backtested with available daily prices and point-in-time fundamentals, but the China results cannot be replicated at scale.

   original paper

8 more papers written up in this period. Open the Quant Paper Radar →