Research map / Label-efficient and robust learning / Semi-supervised learning
Learning the Right Layers a Data-Driven Layer-Aggregation Strategy for Semi-Supervised Learning on Multilayer Graphs
Sara Venturini, Andrea Cristofari, Francesco Rinaldi, Francesco Tudisco
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 |
|---|---|---|---|
| learned generalized mean aggregation with bilevel Frank-Wolfe (BINOM vs MULTI) | multilayer SSL on synthetic and real datasets · synthetic and real multilayer graphs | APR | Improved |
| learned aggregation (BINOM) versus fixed aggregations | real datasets, 1% labels · real-world multilayer datasets with one noisy layer | APR: 0.97 vs 0.94 for best fixed mean (ARIT) | Improved |
| fixed min aggregation | multilayer SSL · synthetic all settings | accuracy | Hurt |
| adding a noisy layer to real datasets with fixed max aggregation | MAX aggregation · UCI 1% labels | accuracy: 0.85 without noise vs 0.5 with one noise layer | Hurt |
| fixed arithmetic/harmonic/max aggregation | multilayer SSL · synthetic noisy case | accuracy: MAX 0.68 vs learned 0.99 at std=2 | Hurt |
| AGML and SGMI baselines | multilayer SSL · real datasets 10% labels | APR: SGMI APR 0.85 vs BINOM 0.98 | Hurt |
| learned aggregation (BINOM/MULTI) versus single layers | real and synthetic datasets · real-world datasets | accuracy | Improved |
| additional layers of noise (one to two) added to real datasets | fixed ARIT aggregation · 3sources 1% labels | accuracy: 0.79 clean vs 0.73 (+1 noise) vs 0.70 (+2 noise) | Hurt |
Closest papers by meaning
- Community Extraction in Multilayer Networks with Heterogeneous Community Structure (JMLR 2017 · similarity 0.88)
- Semi-supervised Community Detection via Structural Similarity Metrics (ICLR 2023 · similarity 0.87)
- Continuous Partitioning for Graph-Based Semi-Supervised Learning (NeurIPS 2024 · similarity 0.86)
- Unsupervised Neighborhood Propagation Kernel Layers for Semi-supervised Node Classification (AAAI 2024 · similarity 0.86)
- Revisiting Self-Supervised Heterogeneous Graph Learning from Spectral Clustering Perspective (NeurIPS 2024 · similarity 0.86)
- Local Graph Clustering with Noisy Labels (ICLR 2024 · similarity 0.86)
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.
