Research map / Probabilistic and scientific modeling / Variational and generative sampling
Differentiable Random Partition Models
Thomas M. Sutter, Alain Ryser, Joram Liebeskind, Julia E Vogt
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 |
|---|---|---|---|
| model dependent cluster assignments with DRPM instead of i.i.d. categorical (VaDE) | DRPM-VC · MNIST | ACC: 0.82 -> 0.94 | Improved |
| model dependent cluster assignments with DRPM instead of i.i.d. categorical (VaDE) | DRPM-VC · FMNIST | ACC: 0.56 -> 0.65 | Improved |
| DRPM-VC versus GMM in latent space on MNIST | DRPM-VC · MNIST | NMI: 0.86 -> 0.89 | Improved |
| use pretrained encoder embeddings instead of raw data space for GMM | GMM · MNIST | ACC: 0.41 -> 0.88 | Improved |
| DRPM-VAE learns assignment of latent dimensions to shared factors instead of HG-VAE heuristic | DRPM-VAE · mpi3d, n_s=1 | shared-factor balanced accuracy: 0.22 -> 0.39 | Improved |
| DRPM-VAE versus HG-VAE | DRPM-VAE · mpi3d, n_s=5 | shared-factor balanced accuracy: 0.28 -> 0.42 | Improved |
| DRPM-VAE versus HG-VAE on independent factors at n_s=5 | DRPM-VAE · mpi3d, n_s=5 | independent-factor balanced accuracy: 0.01 -> 0.01 | No clear effect |
| DRPM-VAE versus HG-VAE on estimating number of shared factors | DRPM-VAE · mpi3d | MSE of estimated n_s | No clear effect |
| DRPM-MTL partitions last shared layer by task instead of unitary loss scaling | DRPM-MTL · noisyMultiMNIST | task accuracy | Mixed |
| increase noise ratio alpha on the right task from 0.6 to 0.8 | DRPM-MTL · noisyMultiMNIST | dimensions assigned to right task | Improved |
| maximum noise ratio alpha=0.9 | DRPM-MTL · noisyMultiMNIST | average accuracy | Hurt |
Closest papers by meaning
- Neural Variational Inference and Learning in Undirected Graphical Models (NeurIPS 2017 · similarity 0.87)
- Recursive Inference for Variational Autoencoders (NeurIPS 2020 · similarity 0.87)
- Monotonic Alpha-divergence Minimisation for Variational Inference (JMLR 2023 · similarity 0.87)
- Learning Markov Random Fields for Combinatorial Structures via Sampling through Lovász Local Lemma (AAAI 2023 · similarity 0.86)
- Meta-Amortized Variational Inference and Learning (AAAI 2020 · similarity 0.86)
- Joint Learning of Energy-based Models and their Partition Function (ICML 2025 · 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.
