Low-Overhead Quantum Error Correction Codes for Noisy Intermediate-Scale Quantum Devices

¹ Department of CSE, School of Technology, The Apollo University, Chittoor, Andhra Pradesh, India
² Department of IT department, S V College of Engineering, Andhra Pradesh, India
³ Department of CSE, Sree Rama Engineering College (Autonomous), Andhra Pradesh, India
⁴ Department of Computer Science and Engineering, Vel Tech Rangarajan Dr. Sagunthala R&D Institute of Science and Technology, Avadi, Chennai, India.

^* Corresponding author: This email address is being protected from spambots. You need JavaScript enabled to view it.

Published online: 23 March 2026

Abstract

Noisy Intermediate-Scale Quantum (NISQ) devices represent current quantum computing technology with 50-1000 qubits operating without comprehensive fault-tolerant error correction. The fundamental challenge lies in balancing error correction necessity against prohibitive resource overhead. Conventional quantum error correction codes require 10:1 to 1000:1 qubit ratios and syndrome extraction circuits exceeding 50 gates, consuming entire NISQ capacities. This work presents a comprehensive low-overhead quantum error correction framework specifically designed for NISQ constraints. Our approach combines three elements: optimised stabiliser codes which get qubit ratios of 3:1 to 10:1, adaptive syndrome extraction which works with circuits with less than 20 gates and machine learning-enhanced decoding which is specific to the noise profiles of the hardware. Experimental validation on IBM Quantum (127- qubit Eagle), IonQ trapped-ion systems, and Google Sycamore simulation demonstrates successful quantum state protection achieving 2.7× error suppression with 60% overhead reduction versus conventional surface codes. The proposed code achieves logical error rates of 0.0065 using only 10 physical qubits compared to 25 qubits for equivalent surface codes. Machine learning decoder attains 94.3% accuracy with 12-microsecond inference latency, enabling real-time error correction. Break-even performance achieved at 0.8% physical error rates establishes practical pathways for near-term quantum advantage in variational algorithms and quantum simulation applications.