Detection of noise correlations in two qubit systems by machine learning
Detection of noise correlations in two qubit systems by machine learning
Authors: Dario Fasone, Shreyasi Mukherjee, Dario Penna, Fabio CirinnĂ , Mauro Paternostro, Elisabetta Paladino, Luigi Giannelli, G. Falci Date: 2026-04-21 Paper ID: openalex:2509.03389
Summary
This paper proposes a machine learning-based quantum sensing protocol to detect and classify different types of noise in two-qubit systems. By leveraging the sensitivity of coherent population transfer efficiencies under specific driving conditions, the method distinguishes between six classes of Markovian and non-Markovian noise with high accuracy (>94%). The approach is notable for its efficiency, as it circumvents the need for real-time monitoring or time-series data collection by utilizing a shallow neural network. This framework demonstrates how machine learning can enhance the characterization of quantum hardware non-idealities.
Key Contributions
- Introduces a machine-learning assisted quantum sensing protocol for classifying spatial and temporal correlations of classical noise in two-qubit systems.
- Achieves >94% classification accuracy for six distinct Markovian and non-Markovian noise classes using only final transfer efficiencies from a coherent population transfer protocol.
- Reduces experimental resource requirements by avoiding time-series measurements, relying instead on a shallow neural network trained on inputs from specific driving conditions.
Links
Metadata & Links
- url
- https://arxiv.org/abs/2509.03389
- paper_id
- 2509.03389
- paper_source
- openalex
- domain
- physics-ml
- tags
- quantum-sensingnoise-characterizationmachine-learning-in-physicstwo-qubit-systemsnon-markovian-noise
- architectures
-
- datasets
-
- concept_slugs
-
- dataset_slugs
-
- skill
- TimeSeriesSkill
- created_at
- 2026-04-24T05:52:20Z