Research map / Probabilistic and scientific modeling / Variational and generative sampling
Towards K-means-friendly Spaces: Simultaneous Deep Learning and Clustering
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 |
|---|---|---|---|
| jointly optimize K-means structure with deep dimensionality reduction rather than train SAE then apply K-means | DCN · RCV1-v2, 4 Clust. | NMI: 0.73 -> 0.76 | Helped |
| jointly optimize K-means structure with deep dimensionality reduction rather than train SAE then apply K-means | DCN · 20Newsgroup | ARI: 0.28 -> 0.34 | Helped |
| jointly optimize K-means structure with deep dimensionality reduction rather than train SAE then apply K-means | DCN · Raw MNIST | NMI: 0.80 -> 0.81 | Helped |
| jointly optimize K-means structure with deep dimensionality reduction rather than train SAE then apply K-means | DCN · Pre-processed MNIST | ACC: 0.93 -> 0.95 | Helped |
| jointly optimize K-means structure with deep dimensionality reduction rather than train SAE then apply K-means | DCN · Pendigits | NMI: 0.65 -> 0.69 | Helped |
| remove the reconstruction term from the jointly trained deep clustering objective (ablation) | DCN · Synthetic nonlinear data | K-means clustering suitability | Hurt |
| use SAE training without K-means regularization instead of the DCN objective (ablation) | SAE+KM · MNIST | clustering performance | Hurt |
| increase the clustering weight λ beyond an appropriate range (ablation) | DCN · MNIST | clustering performance | Hurt |
| use layer-wise pre-training to initialize the network (ablation) | DCN · Unsupervised clustering | optimization effectiveness | Helped |
| apply DCN instead of SAE+KM on 20Newsgroup | DCN · 20Newsgroup | NMI: 0.47 -> 0.48 | Helped |
Closest papers by meaning
- Unsupervised Deep Embedding for Clustering Analysis (ICML 2016 · similarity 0.88)
- DeepDPM: Deep Clustering With an Unknown Number of Clusters (CVPR 2022 · similarity 0.88)
- Efficient Deep Embedded Subspace Clustering (CVPR 2022 · similarity 0.88)
- Deep Subspace Clustering Networks (NeurIPS 2017 · similarity 0.88)
- Exploring a Principled Framework for Deep Subspace Clustering (ICLR 2025 · similarity 0.87)
- You Can Trust Your Clustering Model: A Parameter-free Self-Boosting Plug-in for Deep Clustering (NeurIPS 2025 · similarity 0.86)
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.