Quantum Attack Dataset (QAD)
- Post by: Pasindu Udugahapattuwa, Madhusanka Liyanage, Chamitha De Alwis and Uditha Wijewardhana
- March 23, 2026
- No Comment
The Quantum Attack Dataset (QAD) is a comprehensive synthetic dataset developed to support research on quantum attack detection using Artificial Intelligence. It was created as part of ongoing research at University College Dublin, Ireland and University of Sri Jayewardenepura, Sri Lanka, specifically to address the quantum attack scenarios by concerning balanced and unbalanced frequencies using simulation-based machine learning–ready data to evaluate quantum attacks in various scenarios.
Unlike existing quantum datasets that typically focus on attacks such as Gate tampering, Noise models, Measurement tampering, and Pulse-level distortions spans two dimensions of frequencies:
- Balanced dataset- seven quantum attack scenarios considering equal frequencies with clean benign dataset
- Unbalanced dataset- seven quantum attack scenarios considering unequal frequencies with clean benign dataset
The dataset includes both benign and attack scenarios, with adversarial frequencies modeled using simulated quantum circuit environment. Simulated attacks include Gate tampering, Noise models, Measurement tampering, Pulse-level distortions, among others. To label ambiguous benign data at scale, Qiskit simulation was applied — generating quantum attack data using baseline circuit and quantum attack dispatcher extracting quantum attack features.
Quantum Attack Dataset Generation is a configurable pipeline that produces large-scale, labeled datasets of quantum attack scenarios. It builds parametrized quantum circuits representing both benign and adversarial operations, executes them on Qiskit simulators with realistic NISQ-inspired noise models, and records detailed metadata such as circuit parameters, noise settings, and attack labels. By combining templated circuits with diverse attack and noise configurations, the framework supports reproducible benchmarking, quantum attack detection, and resilience testing, offering a scalable foundation for developing and evaluating quantum security and machine learning models.
Download the dataset from Kaggle: A Comprehensive Quantum Attack Dataset (QAD)
DOI: https://doi.org/10.34740/kaggle/ds/9778273
This dataset serves as a valuable resource for researchers and academics working on quantum cybersecurity, adversarial modeling, and quantum trust management approaches.
