Research map / Probabilistic and scientific modeling / Molecules and proteins
Recurrent Quantum Neural Networks
Johannes Bausch
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Use UMAP rather than PCA plus t-SNE preprocessing | QRNN on full MNIST · Full MNIST | Accuracy [%]: $94.6\pm0.4$ -> $96.7\pm0.2$ | Improved |
| Ensemble 3 UMAP-preprocessed QRNNs | QRNN on full MNIST · Full MNIST | Accuracy [%]: $96.7\pm0.2$ -> $\mathbf{99.23\pm0.05}$ | Improved |
| Ensemble 4 QRNNs | QRNN on MNIST {0, 1} · MNIST {0, 1} | Accuracy [%]: $99.2\pm0.2$ -> $\mathbf{99.6\pm0.16}$ | Improved |
| Use a 10-qubit L-BFGS QRNN rather than a 12-qubit Adam QRNN | QRNN on MNIST {3, 6} · MNIST {3, 6} | Accuracy [%]: $89.7\pm 0.8$ -> $97.1\pm0.7$ | Improved |
| Ensemble 6 QRNNs | QRNN on MNIST {3, 6} · MNIST {3, 6} | Accuracy [%]: $97.1\pm0.7$ -> $\mathbf{99.0\pm0.3}$ | Improved |
| Use UMAP preprocessing rather than pixel-by-pixel inputs | QRNN on full MNIST · Full MNIST | Accuracy [%]: $96.7\pm0.2$ vs. $94.6\pm0.4$ | Improved |
| Use UMAP instead of t-SNE as the dimensionality reduction method | QRNN preprocessing for MNIST · Full MNIST | Accuracy [%] | Improved |
| Use Adam rather than L-BFGS for general QRNN training | QRNN · Sequence learning tasks | Convergence reliability | Improved |
| Use a neuron order of 2 rather than other tested orders (ablation) | QRNN sentence-learning topology · Sentence structure learning | Training steps to convergence | Improved |
| Use degree 2 rather than degree 3 or 4 neurons (ablation) | QRNN sentence-learning topology · Sentence structure learning | Training steps to convergence | Improved |
| Use two work stages rather than one (ablation) | QRNN sentence-learning topology · Sentence structure learning | Training steps to convergence | Improved |
| Increase work stages beyond two (ablation) | QRNN sentence-learning topology · Sentence structure learning | Training steps to convergence | Hurt |
Closest papers by meaning
- Quantum Gated Recurrent Neural Networks (TPAMI 2024 · similarity 0.86)
- Feedback-driven recurrent quantum neural network universality (ICLR 2026 · similarity 0.85)
- Eigenvalue Normalized Recurrent Neural Networks for Short Term Memory (AAAI 2020 · similarity 0.84)
- Full-Capacity Unitary Recurrent Neural Networks (NeurIPS 2016 · similarity 0.83)
- Unitary Evolution Recurrent Neural Networks (ICML 2016 · similarity 0.82)
- Learning Useful Representations of Recurrent Neural Network Weight Matrices (ICML 2024 · similarity 0.81)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
