Learning to Program Variational Quantum Circuits with Fast Weights
Samuel Yen-Chi Chen
TL;DR
The paper tackles the slow, gradient-heavy training of quantum recurrent architectures for time-series and reinforcement learning by introducing Quantum Fast Weight Programmers (QFWP), a hybrid scheme where a slow classical network generates additive updates to a fast variational quantum circuit. This slow-fast programming approach preserves memory across time steps without explicit quantum recurrence, enabling efficient learning on NISQ-like hardware. Through extensive time-series benchmarks (Damped SHM, Bessel J2, NARMA5/10) and MiniGrid RL with asynchronous QA3C, QFWP matches or exceeds the performance of the quantum LSTM baselines while using smaller quantum models and faster convergence, and often exhibits greater stability. The work demonstrates a viable path for hybrid quantum-classical sequential learning that sidesteps the computational burdens of backpropagation-through-time in QRNNs, highlighting practical potential for scalable quantum-assisted control and forecasting.
Abstract
Quantum Machine Learning (QML) has surfaced as a pioneering framework addressing sequential control tasks and time-series modeling. It has demonstrated empirical quantum advantages notably within domains such as Reinforcement Learning (RL) and time-series prediction. A significant advancement lies in Quantum Recurrent Neural Networks (QRNNs), specifically tailored for memory-intensive tasks encompassing partially observable environments and non-linear time-series prediction. Nevertheless, QRNN-based models encounter challenges, notably prolonged training duration stemming from the necessity to compute quantum gradients using backpropagation-through-time (BPTT). This predicament exacerbates when executing the complete model on quantum devices, primarily due to the substantial demand for circuit evaluation arising from the parameter-shift rule. This paper introduces the Quantum Fast Weight Programmers (QFWP) as a solution to the temporal or sequential learning challenge. The QFWP leverages a classical neural network (referred to as the 'slow programmer') functioning as a quantum programmer to swiftly modify the parameters of a variational quantum circuit (termed the 'fast programmer'). Instead of completely overwriting the fast programmer at each time-step, the slow programmer generates parameter changes or updates for the quantum circuit parameters. This approach enables the fast programmer to incorporate past observations or information. Notably, the proposed QFWP model achieves learning of temporal dependencies without necessitating the use of quantum recurrent neural networks. Numerical simulations conducted in this study showcase the efficacy of the proposed QFWP model in both time-series prediction and RL tasks. The model exhibits performance levels either comparable to or surpassing those achieved by QLSTM-based models.