Research map / Probabilistic and scientific modeling / Variational and generative sampling
Laplacian Autoencoders for Learning Stochastic Representations
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 |
|---|---|---|---|
| train the Laplace approximation online rather than post-hoc | Laplacian autoencoder · Hessian approximation experiment | - log p(x): 283.3 ± 88.6 -> 25.8 ± 0.2 | Helped |
| use approximate diagonal rather than exact diagonal Hessian (ablation) | Laplacian autoencoder · Hessian approximation experiment | MSE: 25.7 ± 0.2 -> 25.8 ± 0.4 | No clear effect |
| use approximate diagonal rather than exact diagonal Hessian (ablation) | Laplacian autoencoder · Hessian approximation experiment | - log p(x): 25.8 ± 0.2 -> 25.9 ± 0.4 | No clear effect |
| use online LAE instead of AE Ensemble | MNIST data imputation · MNIST data imputation | MCE: 0.2 -> 0.16 | Helped |
| use stochastic online LAE representations instead of deterministic AE representations | Semi-supervised classifier · MNIST, few labels per class (1-20) | accuracy | Helped |
| use stochastic online LAE representations instead of AE representations | Semi-supervised classifier · MNIST, more than 100 labels per class | accuracy | Hurt |
| use online LAE instead of deterministic AE | Semi-supervised classifier · CelebA (10 labeled datapoints) | accuracy: 0.73 -> 0.74 | No clear effect |
Closest papers by meaning
- Stick-Breaking Variational Autoencoders (ICLR 2017 · similarity 0.90)
- Model Selection for Bayesian Autoencoders (NeurIPS 2021 · similarity 0.89)
- Variational Laplace Autoencoders (ICML 2019 · similarity 0.89)
- Variational Learning of Fractional Posteriors (ICML 2025 · similarity 0.88)
- Shape your Space: A Gaussian Mixture Regularization Approach to Deterministic Autoencoders (NeurIPS 2021 · similarity 0.88)
- A Statistical Analysis of Wasserstein Autoencoders for Intrinsically Low-dimensional Data (ICLR 2024 · similarity 0.87)
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.