QMetric: Benchmarking Quantum Neural Networks Across Circuits, Features, and Training Dimensions
Silvie Illésová, Tomasz Rybotycki, Martin Beseda · 2025
Read QMetric: Benchmarking Quantum Neural Networks Across Circuits, Features, and Training Dimensions by Silvie Illésová, Tomasz Rybotycki, Martin Beseda (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2506-23765.txt.
Description
Silvie Illésová, Tomasz Rybotycki, and Martin Beseda present QMetric, a Python package for judging quantum neural networks on more than raw accuracy. Hybrid quantum-classical models are gaining ground in machine learning, yet builders lack tools to see why one works or fails. QMetric offers a suite of readable metrics. They cover how expressive a quantum circuit is, how its features are represented, and how training behaves over time.
The measures include circuit fidelity, entanglement entropy, the risk of barren plateaus where training stalls, and training stability. The package works with Qiskit and PyTorch, two standard toolkits. A case study on sorting handwritten digits compares classical and quantum-enhanced models side by side. Code, plots, and a reproducible setup are shared in public. The authors aim to make quantum model testing more open and more informative.
Catalog record
| CREATOR | Silvie Illésová, Tomasz Rybotycki, Martin Beseda |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 11 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2506-23765