Research map / Detection and segmentation / Open-set and few-shot detection
Greedy Structure Learning of Hierarchical Compositional Models
Adam Kortylewski, Aleksander Wieczorek, Mario Wieser, Clemens Blumer, Sonali Parbhoo, Andreas Morel-Forster, Volker Roth, Thomas Vetter
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 the hierarchical structure instead of specifying it a-priori | OURS versus HABM · Four Domain unsupervised domain adaptation | classification score: OURS scores are higher in six out of eight experiments | Improved |
| learn the hierarchical structure instead of specifying it a-priori | OURS versus HABM · Four Domain unsupervised, D → A and D → W | classification score: 34.6 $\pm$ 3.7 vs 33.1 $\pm$ 2.4; 68.2 $\pm$ 2.9 vs 65.6 $\pm$ 3.8 | Hurt |
| learn the hierarchical structure instead of specifying it a-priori | OURS versus other methods · Four Domain semi-supervised domain adaptation | classification score: OURS scores are higher in five out of eight experiments | Improved |
| compose two parts at each hierarchy layer rather than changing the number composed (ablation) | greedy structure learning scheme · Four Domain classification experiments | classification score | No clear effect |
| use a single holistic object model instead of multiple templates | OURS · Four Domain unsupervised, DSLR source | classification score | Hurt |
Closest papers by meaning
- Inducing Hierarchical Compositional Model by Sparsifying Generator Network (CVPR 2020 · similarity 0.88)
- Learning Hierarchical Image Segmentation For Recognition and By Recognition (ICLR 2024 · similarity 0.87)
- Compositional Scene Understanding through Inverse Generative Modeling (ICML 2025 · similarity 0.86)
- Learning to Compose: Improving Object Centric Learning by Injecting Compositionality (ICLR 2024 · similarity 0.85)
- Learning Partonomic 3D Reconstruction from Image Collections (CVPR 2025 · similarity 0.85)
- Hierarchy-Agnostic Unsupervised Segmentation: Parsing Semantic Image Structure (NeurIPS 2024 · 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.
