Research map / Probabilistic and scientific modeling / Variational and generative sampling
Missing Data Imputation and Acquisition with Deep Hierarchical Models and Hamiltonian Monte Carlo
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| add hierarchical latent layers (VAE to H-VAE) (ablation) | VAE · MNIST | NLL of unobserved features: 0.124 -> 0.119 | Helped |
| add HMC inference (VAE to HMC-VAE) (ablation) | VAE · MNIST | NLL of unobserved features: 0.124 -> 0.101 | Helped |
| combine hierarchy and HMC (HH-VAE) (ablation) | VAE · MNIST | NLL of unobserved features: 0.124 -> 0.094 | Helped |
| combine hierarchy and HMC (HH-VAE) (ablation) | VAE · Fashion-MNIST | NLL of unobserved features: 0.162 -> 0.144 | Helped |
| add HMC and hierarchy (ablation) | VAE · MNIST | NLL of predicted target: 0.153 -> 0.056 | Helped |
| add HMC and hierarchy (ablation) | VAE · Fashion-MNIST | NLL of predicted target: 0.501 -> 0.337 | Helped |
| add hierarchical layers without HMC (H-VAE) (ablation) | VAE · MNIST | accuracy of predicted digits: 0.953 -> 0.953 | No clear effect |
| add HMC and hierarchy (ablation) | VAE · MNIST | accuracy of predicted digits: 0.953 -> 0.981 | Helped |
| add HMC and hierarchy (ablation) | VAE · Fashion-MNIST | accuracy of predicted digits: 0.824 -> 0.876 | Helped |
| replace VAEM with HH-VAEM | VAEM · UCI Yatch | RMSE of unobserved features: 0.93 -> 0.90 | Helped |
| replace HMC-VAE with HH-VAEM (ablation) | HMC-VAE · UCI Insurance | NLL of observed features: 0.92 -> 0.95 | Hurt |
| HH-VAEM replaces VAEM | VAEM · UCI Bank | training time (minutes): 29.92 -> 316.81 | Hurt |
Closest papers by meaning
- DiffPuter: Empowering Diffusion Models for Missing Data Imputation (ICLR 2025 · similarity 0.87)
- Learning What Not to Impute: An Uncertainty-Aware Diffusion Framework for Meaningful Missingness (NeurIPS 2026 · similarity 0.86)
- Missing Data Imputation by Reducing Mutual Information with Rectified Flows (NeurIPS 2025 · similarity 0.86)
- Rethinking the Diffusion Models for Missing Data Imputation: A Gradient Flow Perspective (NeurIPS 2024 · similarity 0.85)
- Latent variable model for high-dimensional point process with structured missingness (ICML 2024 · similarity 0.85)
- Generative Modeling under Non-Monotone MAR Missingness via Approximate Wasserstein Gradient Flows (NeurIPS 2026 · similarity 0.85)
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.