Attention-Based Deep Reinforcement Learning for Qubit Allocation in Modular Quantum Architectures
Enrico Russo, Maurizio Palesi, Davide Patti, Giuseppe Ascia, Vincenzo Catania
TL;DR
This work tackles the NP-hard problem of qubit allocation in modular quantum architectures by framing it as a sequential, constraint-aware DRL task. It introduces a Transformer-based policy with Graph Neural Network encoders and a core snapshot mechanism to produce feasible, low-inter-core-transfer allocations across circuit slices, employing a masked attention decoder and REINFORCE training. Empirical results show the learned policy reduces inter-core communications and improves online time-to-solution relative to derivative-free baselines and competitive state-of-the-art methods on grid and all-to-all architectures, with some limitations on highly structured circuits. The approach advances scalable quantum compilation by providing a near-term, learnable heuristic capable of generalizing to varying circuit depths and qubit counts, and suggests avenues for richer masking, PPO-based training, and diversified circuit datasets.
Abstract
Modular, distributed and multi-core architectures are currently considered a promising approach for scalability of quantum computing systems. The integration of multiple Quantum Processing Units necessitates classical and quantum-coherent communication, introducing challenges related to noise and quantum decoherence in quantum state transfers between cores. Optimizing communication becomes imperative, and the compilation and mapping of quantum circuits onto physical qubits must minimize state transfers while adhering to architectural constraints. The compilation process, inherently an NP-hard problem, demands extensive search times even with a small number of qubits to be solved to optimality. To address this challenge efficiently, we advocate for the utilization of heuristic mappers that can rapidly generate solutions. In this work, we propose a novel approach employing Deep Reinforcement Learning (DRL) methods to learn these heuristics for a specific multi-core architecture. Our DRL agent incorporates a Transformer encoder and Graph Neural Networks. It encodes quantum circuits using self-attention mechanisms and produce outputs through an attention-based pointer mechanism that directly signifies the probability of matching logical qubits with physical cores. This enables the selection of optimal cores for logical qubits efficiently. Experimental evaluations show that the proposed method can outperform baseline approaches in terms of reducing inter-core communications and minimizing online time-to-solution. This research contributes to the advancement of scalable quantum computing systems by introducing a novel learning-based heuristic approach for efficient quantum circuit compilation and mapping.