# Attention-Based Deep Reinforcement Learning for Qubit Allocation in Modular Quantum Architectures Source: [Attention-Based Deep Reinforcement Learning for Qubit Allocation in Modular Quantum Architectures]() Collection index: [llms.txt]() Data updated: 2026-04-29T05:00:12.215Z Authors: Enrico Russo, Maurizio Palesi, Davide Patti, Giuseppe Ascia, Vincenzo Catania Published: 2024-06-17T12:09:11.000Z ## 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. ## Recorded benchmark results A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. - [Circuit Allocation on QNN 50 qubits](): Circuit Allocation; QNN 50 qubits; Inter-core Communication Overhead: 573. - [Circuit Allocation on Random Avg 50 qubits](): Circuit Allocation; Random Avg 50 qubits; Inter-core Communications: 208. - [Circuit Allocation on Cuccaro Adder 50 qubits](): Circuit Allocation; Cuccaro Adder 50 qubits; Inter-core Communications: 143. - [Circuit Allocation on Deutsch Jozsa 50 qubits](): Circuit Allocation; Deutsch Jozsa 50 qubits; Inter-core Communications: 48. - [Circuit Allocation on Draper Adder 50 qubits](): Circuit Allocation; Draper Adder 50 qubits; Inter-core Communications: 691. - [Circuit Allocation on Graph State 50 qubits](): Circuit Allocation; Graph State 50 qubits; Inter-core Communication: 546. - [Circuit Allocation on QFT 50 qubits](): Circuit Allocation; QFT 50 qubits; Inter-core Communications: 1146. - [Circuit Allocation on Quantum Volume 50 qubits](): Circuit Allocation; Quantum Volume 50 qubits; Inter-core Communication Overhead: 1204. - [Circuit Allocation on Cuccaro Adder 100 Qubits](): Circuit Allocation; Cuccaro Adder 100 Qubits; Inter-core Communications: 517. - [Circuit Allocation on Deutsch Jozsa 100 Qubits](): Circuit Allocation; Deutsch Jozsa 100 Qubits; Inter-core Communication Overhead: 363. - [Circuit Allocation on Draper Adder 100 Qubits](): Circuit Allocation; Draper Adder 100 Qubits; Inter-core Communications: 2631. - [Circuit Allocation on Graph State 100 Qubits](): Circuit Allocation; Graph State 100 Qubits; Inter-core Communication Overhead: 2892. - [Circuit Allocation on QFT 100 Qubits](): Circuit Allocation; QFT 100 Qubits; Inter-core Communications: 4887. - [Circuit Allocation on QNN 100 Qubits](): Circuit Allocation; QNN 100 Qubits; Inter-core Comms.: 3549. - [Circuit Allocation on Quantum Volume 100 Qubits](): Circuit Allocation; Quantum Volume 100 Qubits; Inter-core Communication Overhead: 5289. - [Circuit Allocation on Random Avg 100 Qubits](): Circuit Allocation; Random Avg 100 Qubits; Inter-core Communication Overhead: 500. ## Original sources and implementation [Original paper]() [PDF]()