Research map / Probabilistic and scientific modeling / Variational and generative sampling
Quantum latent distributions in deep generative models
Omar Bacarreza, Thorin Farnsworth, Alexander Makarovskiy, Hugo Wallner, Tessa Hicks, Santiago Sempere-Llagostera, John J Price, Robert J. A. Francis-Jones, William R. Clements
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 quantum instead of distinguishable-photon latent (ablation) | GAN · Quantum dataset (synthetic) | L1 distance: 0.041 -> 0.036 | Improved |
| use quantum instead of Gaussian latent | GAN · Quantum dataset (synthetic) | L1 distance: 0.061 -> 0.036 | Improved |
| use quantum latent instead of Gaussian | GAN · Bernoulli dataset (synthetic) | L1 distance: 0.012 -> 0.015 | No clear effect |
| use quantum instead of distinguishable-photon latent (ablation) | MolGAN-based GAN · QM9 d_z=16 | FCD: 1.333 -> 1.160 | Improved |
| use quantum instead of Gaussian latent | MolGAN-based GAN · QM9 d_z=16 | # valid and unique: 1814 -> 2522 | Improved |
| use quantum instead of distinguishable-photon latent (ablation) | MolGAN-based GAN · QM9 d_z=32 | FCD: 1.533 -> 1.536 | No clear effect |
| use quantum instead of distinguishable-photon latent (ablation) | MolGAN-based GAN · QM9 d_z=48 | # valid and unique: 1307 -> 1528 | Improved |
| use quantum instead of Bernoulli latent | MolGAN-based GAN · QM9 d_z=48 | FCD: 1.671 -> 1.696 | No clear effect |
| increase latent size from 16 to 48 with quantum latent | MolGAN-based GAN · QM9 | FCD: 1.160 -> 1.696 | Hurt |
| use quantum 1-1 circuit instead of distinguishable photons (ablation) | MolGAN-based GAN · QM9 d_z=16 (1-1) | FCD: 1.291 -> 1.245 | Improved |
| use quantum instead of distinguishable photons in 1-3-9 circuit (ablation) | MolGAN-based GAN · QM9 d_z=16 | FCD: 1.376 -> 1.187 | Improved |
| use real quantum processor instead of simulation | MolGAN-based GAN · QM9 d_z=32 | FCD: 1.414 -> 1.410 | No clear effect |
Closest papers by meaning
- Characterizing Trainability of Instantaneous Quantum Polynomial Circuit Born Machine (NeurIPS 2026 · similarity 0.86)
- Rethink the Role of Deep Learning towards Large-scale Quantum Systems (ICML 2025 · similarity 0.84)
- Statistically Optimal Generative Modeling with Maximum Deviation from the Empirical Distribution (ICML 2024 · similarity 0.84)
- Understanding Diffusion Models by Feynman's Path Integral (ICML 2024 · similarity 0.84)
- Predictive Performance of Deep Quantum Data Re-uploading Models (ICML 2025 · similarity 0.84)
- BridgeTwist: Twisting Schrödinger Bridges for Training-Free Conditional Sampling (NeurIPS 2026 · similarity 0.84)
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.
