Financial Innovation· 2026Q1
AlphaZeroBeta: deep reinforcement learning for market-neutral portfolios
- 0citations
- Q1SCImago
- 2026year
Short summary
AlphaZeroBeta, a deep reinforcement learning framework, achieves higher Sharpe ratios and near-zero market beta by combining a composite reward function with a CNN-GRU policy trained via Recurrent PPO.
AI-generated from the title and abstract; the full text is not read.
Key points
- Proposes AlphaZeroBeta, a deep reinforcement learning framework for market-neutral portfolios.
- Utilizes a composite reward function balancing excess return, benchmark correlation, and transaction costs.
- Employs a CNN-GRU policy trained with Recurrent PPO.
- Backtests show higher Sharpe ratios and near-zero benchmark correlations than baselines (2014-2024).
AI-generated from the title and abstract; the full text is not read.
Abstract
Abstract Market-neutral portfolios aim to generate consistent returns while offsetting systematic market risk. Traditional approaches based on factor models or convex optimization often underperform during market regime shifts or when structural assumptions break down. We propose AlphaZeroBeta, a deep reinforcement learning framework designed to deliver benchmark-relative alpha (excess returns) with near-zero beta (market neutrality). AlphaZeroBeta combines a composite reward function that balances risk-adjusted excess return, benchmark correlation, and transaction costs, with a CNN-GRU policy trained end-to-end via Recurrent PPO and evaluated through a rolling walk-forward protocol. Backtests covering 2014–2024 across seven equity indices show that the model achieves higher Sharpe ratios than the baselines while maintaining near-zero benchmark correlations and competitive drawdowns.
The authors' abstract, as published at the source. Financial Innovation, 2026 · DOI ↗
The rest is in the Pofolia app
Takeaways and questions to the paper; new summaries every day for your field. Free.
Sign in on the web to openField: Management Science and Operations Research
Management Science and Operations ResearchDecision Sciences