Research map / Probabilistic and scientific modeling / Variational and generative sampling
Multi-Facet Clustering Variational Autoencoders
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 |
|---|---|---|---|
| Use an independent mixture-of-Gaussians prior for each latent facet | MFCVAE · MNIST, 3DShapes, and SVHN | facet disentanglement | Helped |
| Use the ladder architecture with progressive training (ablation) | MFCVAE · Image benchmarks | training stability | Helped |
| Replace the primary ELBO loss form with the alternate loss form (ablation) | Single-facet VaDE-style model · MNIST test | unsupervised clustering accuracy | No clear effect |
| Use MFCVAE instead of ACOL-GAR | SVHN clustering · SVHN test | unsupervised clustering accuracy: 76.80 (1.30) -> 56.25 (0.93) | Hurt |
| Use the style/background facet embedding instead of the digit/shape facet embedding | MFCVAE · MNIST digit class; 3DShapes object shape; SVHN digit class | supervised test accuracy: MNIST: 94.95 (0.04) vs 17.34 (0.24); 3DShapes config. 1: 32.43 (1.38) vs 95.00 (0.45); SVHN: 22.30 (0.16) vs 69.46 (0.36 | Mixed |
| Use the concatenated embeddings from both facets rather than one facet | MFCVAE · MNIST digit class; 3DShapes; SVHN digit class | supervised test accuracy: MNIST: 95.27 (0.07); 3DShapes config. 1: 95.18 (0.42) and 100.00 (0.00); SVHN: 70.39 (0.29) | Helped |
Closest papers by meaning
- Learning Latent Superstructures in Variational Autoencoders for Deep Multidimensional Clustering (ICLR 2019 · similarity 0.92)
- Multi-Level Variational Autoencoder: Learning Disentangled Representations From Grouped Observations (AAAI 2018 · similarity 0.90)
- Tree Variational Autoencoders (NeurIPS 2023 · similarity 0.89)
- Shape your Space: A Gaussian Mixture Regularization Approach to Deterministic Autoencoders (NeurIPS 2021 · similarity 0.89)
- Cooperation in the Latent Space: The Benefits of Adding Mixture Components in Variational Autoencoders (ICML 2023 · similarity 0.89)
- Deep Variational Incomplete Multi-View Clustering: Exploring Shared Clustering Structures (AAAI 2024 · similarity 0.88)
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.