Research map / Label-efficient and robust learning / Semi-supervised learning
Differentiable Clustering with Perturbed Spanning Forests
Lawrence Stewart, Francis Bach, Felipe Llinares-López, Quentin Berthet
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 |
|---|---|---|---|
| learn embeddings through clustering loss instead of cross-entropy classification, with 1000 labels and 6 withheld classes vs baseline with 10,000 labels and all classes | ResNet on CIFAR-10 · CIFAR-10 semi-supervised | clustering error | Improved |
| learn through clustering instead of classification when few labels are available | CNN and ResNet embeddings · MNIST and CIFAR-10 linear probe transfer | downstream classification error | Improved |
| withhold labels for 3 or 6 classes during training (ablation) | clustering-trained model · MNIST and CIFAR-10 | clustering of unseen classes | Mixed |
| learn a linear de-noising transform through the clustering loss | linear model on synthetic 4-cluster data with 2 noise dimensions · synthetic denoising | validation clustering error: converged to zero after 25 gradient batches | Improved |
Closest papers by meaning
- Fairness, Semi-Supervised Learning, and More: A General Framework for Clustering with Stochastic Pairwise Constraints (AAAI 2021 · similarity 0.87)
- Graph-based Clustering Revisited: A Relaxation of Kernel k-Means Perspective (JMLR 2026 · similarity 0.86)
- Approximate Forest Completion and Learning-Augmented Algorithms for Metric Minimum Spanning Trees (ICML 2025 · similarity 0.85)
- Active Clustering with Model-Based Uncertainty Reduction (TPAMI 2016 · similarity 0.85)
- Differentiable Cluster Discovery in Temporal Graphs (NeurIPS 2026 · similarity 0.85)
- A Random Matrix Analysis of Data Stream Clustering: Coping With Limited Memory Resources (ICML 2022 · 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.
