Research map / Label-efficient and robust learning / Self-supervised and contrastive
SI-IGCL: Subject Invariance-aware Inverse Graph Contrastive Learning for Psychiatric Disorder Identification
Jiayu Lu, Yujin Wang, Xiaofeng Liu, Dandan Li, Bin Wang
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 |
|---|---|---|---|
| add hierarchical topology enhanced transformer (HTET) instead of single-level Transformer (ablation) | SI-IGCL · ABIDE | ACC (%): 71.30 -> 75.27 | Improved |
| add reconstruction loss LR alongside inverse contrastive loss (ablation) | SI-IGCL · ABIDE | ACC (%): 51.15 -> 75.27 | Improved |
| add reconstruction loss LR alongside inverse contrastive loss (ablation) | SI-IGCL · ADHD-200 | ACC (%): 66.02 -> 74.26 | Improved |
| add inverse contrastive loss LIC (ablation) | SI-IGCL · ABIDE | ACC (%): 72.35 -> 75.27 | Improved |
| add inverse contrastive loss LIC (ablation) | SI-IGCL · ADHD-200 | ACC (%): 72.21 -> 74.26 | Improved |
| add reconstruction loss to avoid collapse (ablation) | SI-IGCL · ABIDE | average feature variance: 0.0001 -> 0.4309 | Improved |
| add inverse contrastive loss LIC (ablation) | SI-IGCL · ABIDE | nearest-neighbor diversity (lower better): 0.4100 -> 0.3800 | Improved |
| add corrective term alpha to inverse contrastive loss (ablation) | SI-IGCL · ABIDE and ADHD-200 | training loss dynamics | Improved |
| encode FC with subject-invariant encoder | SI-IGCL · ABIDE | subject identification accuracy (%): 98.18 -> 79.21 | Improved |
| use SI-IGCL under leave-one-site-out evaluation | SI-IGCL · ABIDE (leave-one-site-out) | ACC (%): 69.75 -> 71.45 | Improved |
Closest papers by meaning
- Causal Invariance-aware Augmentation for Brain Graph Contrastive Learning (ICML 2025 · similarity 0.88)
- IEBGL:An Interpretability-Enhanced Brain Graph Learning Framework with LLM-Instructed Topology and Literature-Augmented Semantics (CVPR 2026 · similarity 0.85)
- A New Brain Network Construction Paradigm for Brain Disorder via Diffusion-Based Graph Contrastive Learning (TPAMI 2024 · similarity 0.84)
- NeuroTree: Hierarchical Functional Brain Pathway Decoding for Mental Health Disorders (ICML 2025 · similarity 0.84)
- Graph Contrastive Invariant Learning from the Causal Perspective (AAAI 2024 · similarity 0.83)
- MindAligner: Explicit Brain Functional Alignment for Cross-Subject Visual Decoding from Limited fMRI Data (ICML 2025 · similarity 0.82)
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.
