Quantum Adaptive Self-Attention for Financial Rebalancing: An Empirical Study on Automated Market Makers in Decentralized Finance
Quantum Adaptive Self-Attention for Financial Rebalancing: An Empirical Study on Automated Market Makers in Decentralized Finance
Authors: Chi-Sheng Chen, Aidan Hung-Wen Tsai Date: 2026-04-21 Paper ID: openalex:2509.16955
Summary
The paper introduces Quantum Adaptive Self-Attention (QASA), a hybrid quantum-classical module designed for financial time-series decision-making in automated market makers. By constructing query, key, and value vectors through variational quantum circuits (VQCs) and applying softmax attention to Pauli-Z expectation values, QASA integrates quantum processing into transformer architectures. Empirical evaluation on BTCUSDC market data shows that QASA-based models outperform purely classical and pure quantum baselines in risk-adjusted metrics like Return and Sharpe ratio. The study confirms the viability of hybrid quantum-classical attention for complex financial rebalancing tasks.
Key Contributions
- Introduces QASA (Quantum Adaptive Self-Attention), a hybrid quantum-classical attention module utilizing VQCs to process financial time-series data.
- Formulates automated market maker (AMM) rebalancing as a binary detection task, showing QASA achieves superior risk-adjusted returns (13.99%, Sharpe 1.76) compared to classical transformer and pure quantum baselines.
- Provides an empirical comparison on BTCUSDC data demonstrating that QASA-based models offer a favorable performance-stability-cost trade-off over pure quantum or purely classical ensembles.
Open Questions & Future Work
Key Concepts
- quantum-adaptive-self-attention: A hybrid quantum-classical attention mechanism using variational quantum circuits to construct queries, keys, and values from Pauli-Z expectation vectors.
Archivist Review
Approved Quantum Adaptive Self-Attention as a reusable hybrid module. The open question was renamed to be more general while preserving the technical bottleneck regarding hybrid quantum-classical scalability. Standard datasets like BTCUSDC were rejected as they are not unique or novel enough for individual registry.
Approved Concepts
- Quantum Adaptive Self-Attention: Introduces a novel hybrid quantum-classical attention mechanism specifically for financial time-series forecasting.
Approved Open Questions
- Hybrid Quantum Attention Scalability: Understanding the scaling limits is essential for determining if quantum-enhanced financial modules can transition from experimental prototypes to real-world, high-frequency decentralized finance applications.
Rejected Candidates
- [open_question] Scalability of Hybrid QASA Modules (
qasa-scalability-latency-limitations) - duplicate_existing: Renamed to a more canonical slug to reflect the broader mechanism rather than the specific model implementation.
Links
Metadata & Links
- url
- https://arxiv.org/abs/2509.16955
- paper_id
- 2509.16955
- paper_source
- openalex
- domain
- time-series
- tags
-
- architectures
-
- datasets
-
- concept_slugs
- quantum-adaptive-self-attention
- dataset_slugs
-
- skill
- TimeSeriesSkill
- created_at
- 2026-04-24T05:51:49Z