Research map / Probabilistic and scientific modeling / Variational and generative sampling
Decision-Making with Auto-Encoding Variational Bayes
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 |
|---|---|---|---|
| Combine proposals learned by multiple objectives using multiple importance sampling | VAE decision-making procedure · pPCA simulated data | MAE | Helped |
| Use a χ-VAE proposal for a fixed model | pPCA posterior expectation estimation · pPCA simulated data | MAE | Helped |
| Use a Student's t variational distribution instead of a Gaussian (ablation) | χ-VAE · pPCA simulated data | MAE | Helped |
| Use the plugin estimator instead of SNIS | MNIST class-probability estimation · MNIST | classification performance | Helped |
| Use χ-VAE or MIS proposals with the model held fixed | MNIST rejection-based decisions · MNIST | AUPRC | Helped |
| Combine proposals using MIS | MNIST decision-making · MNIST | AUPRC | Helped |
| Train scVI with a VAE objective rather than the other evaluated objectives | scVI · scRNA-seq simulated data | AUPRC: 0.85 versus more than 0.95 | Hurt |
| Use a proposal based on the VAE-trained model | scVI FDR estimation · scRNA-seq simulated data | FDR MAE | Hurt |
| Use the plugin estimator rather than importance sampling with a VAE-based proposal | scVI FDR estimation · scRNA-seq simulated data | FDR MAE | Hurt |
| Combine proposals with MIS | χ-VAE scVI model · scRNA-seq simulated data | FDR MAE | Helped |
| Use AIS instead of the learned MIS proposals | χ-VAE scVI model · scRNA-seq simulated data | FDR MAE | Hurt |
| Increase the number of particles (ablation) | IWAE · pPCA simulated data | held-out likelihood | Hurt |
Closest papers by meaning
- Challenges and Opportunities in High Dimensional Variational Inference (NeurIPS 2021 · similarity 0.87)
- Information Constraints on Auto-Encoding Variational Bayes (NeurIPS 2018 · similarity 0.87)
- Predictive variational inference: Learn the predictively optimal posterior distribution (ICML 2026 · similarity 0.87)
- Variational Learning of Fractional Posteriors (ICML 2025 · similarity 0.86)
- Efficient Mixture Learning in Black-Box Variational Inference (ICML 2024 · similarity 0.86)
- Forward $\chi^2$ Divergence Based Variational Importance Sampling (ICLR 2024 · 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.