Reproduce “Quantum Neural Network Compression” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Zhirui Hu, Peiyan Dong, Zhepeng Wang, Youzuo Lin, Yanzhi Wang, Weiwen Jiang
CompVQC introduces a compilation-aware framework for compressing Variational Quantum Circuits (VQCs), enabling up to 2.57x reduction in transpiled circuit depth. It also significantly improves QNN accuracy and robustness on real noisy quantum hardware, making quantum machine learning more feasible for near-term quantum devices.
Key insights
- Effective quantum neural network compression must be compilation-aware, meaning it accounts for how quantum compilers decompose logical gates into physical operations on target hardware.
- Quantum pruning extends beyond setting parameters to zero; it involves identifying parameter values that cause gates to perform identity operations or introduce only unobservable global phases.
- Quantum quantization is about mapping parameters to values that result in shorter physical circuit depths after compilation, rather than just reducing the numerical precision of parameters.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.