量子记忆机制让投资组合在下跌时防守、上涨时进攻。
Titans-QFWP: A Regime-Aware Hybrid Quantum Fast Weight Programmer for Portfolio Optimization
- 用量子快速调参+记忆模块优化投资策略
- 468只股票测试，年化收益中位数达42.6%
- 适合关注量化投资与量子算法的从业者
我们提出 Titans-QFWP，一种融合量子快速权重编程与钛金风格记忆机制（持久性、意外性、遗忘性）的混合强化学习架构，用于自适应投资组合优化。为应对高维市场特征，引入改进的 A3C^2 框架，结合匈牙利对齐 K-means 聚类和缩放对数收益奖励。在包含约 3,000 个可训练参数的等参数量基准下，对 468 只标普 500 成分股进行评估，该模型表现优异（中位数年化收益率 0.4260，卡玛比率 8.5504，信息比率 0.8427）。消融实验表明，量子门控从根本上重塑了记忆组件的作用：持久性支持回撤控制，意外性促进收益生成，遗忘性提供额外稳定。通过稳定量子表征，模型可在市场下跌时实现防御性配置，同时保留上行潜力。
原文摘要 · Abstract (English)
We propose Titans-QFWP, a hybrid reinforcement learning architecture integrating a Quantum Fast Weight Programmer with Titans-style memory (Persistence, Surprise, and Forgetting) for adaptive portfolio optimization. To address high-dimensional market features, we introduce an enhanced A3C^2 framework with Hungarian-aligned K-means clustering and scaled log-return rewards. Evaluated on 468 S&P 500 stocks under an Equal-Parameter-Count (EPC) benchmark with approximately 3,000 trainable parameters, Titans-QFWP achieves strong performance (median ARR 0.4260, Calmar 8.5504, IR 0.8427). Ablation results reveal that quantum gating fundamentally reshapes memory component roles, with Persistence supporting drawdown control, Surprise contributing to return generation, and Forgetting providing additional stabilization. By stabilizing these quantum representations, the model enables defensive allocation during market drawdowns while preserving upside potential.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。