Research map / Probabilistic and scientific modeling / Variational and generative sampling
Open-Set Recognition with Gaussian Mixture 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 GMVAE embeddings with nearest-centroid uncertainty thresholding instead of CROSR+NC-D | Fashion MNIST open-set recognition · Fashion MNIST withholding 4 classes, Q ≥ 1 | mean-threshold test F1: 3.9% | Helped |
| Use GMVAE embeddings with nearest-centroid uncertainty thresholding instead of CROSR+NC-D | CIFAR-10 open-set recognition · CIFAR-10 withholding 4 classes, K=1, Q ≥ 1 | mean-threshold test F1: 61.3% | Helped |
| Add nearest-centroid uncertainty thresholding to GMVAE with K=(2,2) | MNIST even/odd open-set recognition · MNIST even/odd, Q ≥ 1 | mean-threshold test F1: 19.9% | Helped |
| Add nearest-centroid uncertainty thresholding to GMVAE with K=(2,2) | CIFAR-10 animals/vehicles open-set recognition · CIFAR-10 animals/vehicles, Q ≥ 1 | mean-threshold test F1: 33% | Helped |
| Remove the subcluster v-prior term from the ELBO (ablation) | GMVAE open-set recognition · Open-set experiments | open-set classification performance | Helped |
| Use multiple subclusters per class, K=(2,2), rather than K=1 (ablation) | GMVAE · MNIST even/odd open-set recognition | open-set F1 and robustness to increasing Q | Helped |
| Use GMVAE rather than CROSR for the latent representation with nearest-centroid distance thresholding (ablation) | Open-set recognition · CIFAR-10 withholding 4 classes, K=1 | test F1 | Helped |
| Use CROSR+EVT rather than the simpler nearest-centroid procedures (ablation) | CROSR open-set recognition · Across reported experiments | test F1 | Hurt |
Closest papers by meaning
- Conditional Gaussian Distribution Learning for Open Set Recognition (CVPR 2020 · similarity 0.88)
- C2AE: Class Conditioned Auto-Encoder for Open-Set Recognition (CVPR 2019 · similarity 0.88)
- Generative-Discriminative Feature Representations for Open-Set Recognition (CVPR 2020 · similarity 0.87)
- Classification-Reconstruction Learning for Open-Set Recognition (CVPR 2019 · similarity 0.87)
- Shape your Space: A Gaussian Mixture Regularization Approach to Deterministic Autoencoders (NeurIPS 2021 · similarity 0.87)
- GHOST: Gaussian Hypothesis Open-Set Technique (AAAI 2025 · 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.