Reproduce “Quantum Long Short-Term Memory” 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
Samuel Yen-Chi Chen, Shinjae Yoo, Yao-Lung L. Fang
Brookhaven National Laboratory researchers developed Quantum Long Short-Term Memory (QLSTM), a hybrid quantum-classical model that integrates Variational Quantum Circuits into classical LSTM gates for learning sequential data. Simulations showed QLSTM consistently achieved faster learning, greater accuracy, and more stable training compared to classical LSTMs with similar parameter counts, particularly for complex physical dynamics.
Key insights
- Integrating VQCs allows quantum entanglement to contribute to the model's expressive power, potentially leading to more efficient learning of complex temporal patterns.
- The hybrid variational approach makes the model adaptable for NISQ devices by absorbing noise into tunable parameters, bypassing the need for full quantum error correction.
- Analytical gradient computation via the parameter-shift method enables efficient optimization of the VQC parameters within a classical deep learning framework.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.