Research map / Graph learning / Graph self-supervised learning
Learning Representations without Compositional Assumptions
Tennison Liu, Jeroen Berrevoets, Zhaozhi Qian, Mihaela van der Schaar
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 |
|---|---|---|---|
| Full LEGATO (hierarchical latent graph + learned multi-view graph + orthogonality reg.) vs. best baseline | LEGATO · TCGA | AUROC: 0.703 vs 0.695 (JMVAE) | Improved |
| Remove hierarchical graph pooling (NoHier) (ablation) | LEGATO · TCGA | AUROC: 0.703 vs 0.652 | Improved |
| Replace learned multi-view graph with fully-connected graph (NoGraph) (ablation) | LEGATO · TCGA | AUROC: 0.703 vs 0.696 | Improved |
| Remove orthogonality regularization (NoReg) (ablation) | LEGATO · TCGA | AUROC: 0.703 vs 0.688 | Improved |
| Full LEGATO vs best baseline | LEGATO · UK Biobank | AUROC: 0.720 vs 0.718 (JMVAE) | Improved |
| Remove hierarchical graph pooling (NoHier) (ablation) | LEGATO · UK Biobank | AUROC: 0.720 vs 0.710 | No clear effect |
| Replace learned multi-view graph with fully-connected graph (NoGraph) (ablation) | LEGATO · UK Biobank | AUROC: 0.720 vs 0.698 | Improved |
| Remove orthogonality regularization (NoReg) (ablation) | LEGATO · UK Biobank | AUROC: 0.720 vs 0.679 | Improved |
| Full LEGATO vs baselines on UCI-MFS | LEGATO · UCI-MFS | ACC: 0.824 vs 0.825 (DMVAE, Transformer) | No clear effect |
| Remove hierarchical graph pooling (NoHier) (ablation) | LEGATO · UCI-MFS | ACC: 0.824 vs 0.782 | Improved |
| Replace learned multi-view graph with fully-connected graph (NoGraph) (ablation) | LEGATO · UCI-MFS | ACC: 0.824 vs 0.794 | Improved |
| Remove orthogonality regularization (NoReg) (ablation) | LEGATO · UCI-MFS | ACC: removing it gives 0.801 vs 0.824 full | Hurt |
Closest papers by meaning
- SwitchTab: Switched Autoencoders Are Effective Tabular Learners (AAAI 2024 · similarity 0.87)
- TabGLM: Tabular Graph Language Model for Learning Transferable Representations Through Multi-Modal Consistency Minimization (AAAI 2025 · similarity 0.86)
- ICE-T: Interactions-aware Cross-column Contrastive Embedding for Heterogeneous Tabular Datasets (AAAI 2025 · similarity 0.85)
- A Closer Look at TabPFN v2: Understanding Its Strengths and Extending Its Capabilities (NeurIPS 2025 · similarity 0.85)
- Enhancing Tabular Learners with Context-Aware Semantic Embeddings (NeurIPS 2026 · similarity 0.85)
- Hybrid Autoencoders for Tabular Data: Leveraging Model-Based Augmentation in Low-Label Settings (NeurIPS 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.
