Research map / Label-efficient and robust learning / Recognition and long-tail
Hyperbolic Embedding Inference for Structured Multi-Label Prediction
Bo Xiong, Michael Cochez, Mojtaba Nayyeri, Steffen Staab
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 |
|---|---|---|---|
| remove both implication and exclusion constraints (lambda=0) (ablation) | HMI · CellcycleGO | mAP: 45.58 -> 40.26 | Improved |
| add constraints (HMI vs non constraints) (ablation) | HMI · CellcycleFUN | mAP: 33.87 -> 34.82 | Improved |
| add implication constraints on top of exclusion only (ablation) | HMI · CellcycleGO | mAP: 42.56 -> 45.58 | Improved |
| add exclusion constraints on top of implication only (ablation) | HMI · CellcycleGO | mAP: 44.50 -> 45.58 | Improved |
| add exclusion constraints on top of implication only (ablation) | HMI · CellcycleFUN | HCV: 1.45 -> 1.30 | Improved |
| add implication constraints on top of exclusion only (ablation) | HMI · CellcycleFUN | mAP: 34.70 -> 34.82 | No clear effect |
| penalty weight lambda from 0.0 to 0.005 (ablation) | HMI · CellcycleGO | mAP: 40.26 -> 45.58 | Improved |
| penalty weight lambda from 0.005 to 0.1 (ablation) | HMI · CellcycleGO | mAP: 45.58 -> 41.28 | Hurt |
| increase lambda from 0.0 to 0.1 (ablation) | HMI · CellcycleFUN | HCV: 2.33 -> 0.75 | Improved |
| increase lambda from 0.005 to 0.1 (ablation) | HMI · CellcycleFUN | mAP: 34.82 -> 32.28 | Hurt |
| combine geometric constraints with hyperbolic logistic regression (HMI+HLR vs HLR) (ablation) | HLR · 12 datasets | avg rank mAP: 5.25 -> 2.42 | Improved |
| HMI+HLR instead of HMI (ablation) | HMI · 12 datasets | avg rank mAP: 1.75 -> 2.42 | No clear effect |
Closest papers by meaning
- Hyperbolic Embeddings of Supervised Models (NeurIPS 2024 · similarity 0.90)
- Hyperbolic Entailment Cones for Learning Hierarchical Embeddings (ICML 2018 · similarity 0.88)
- Learning Structured Representations with Hyperbolic Embeddings (NeurIPS 2024 · similarity 0.87)
- Multi-relational Poincaré Graph Embeddings (NeurIPS 2019 · similarity 0.87)
- Smoothing the Geometry of Probabilistic Box Embeddings (ICLR 2019 · similarity 0.87)
- Weighted Embeddings for Low-Dimensional Graph Representation (AAAI 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.
