Research map / Label-efficient and robust learning / Semi-supervised learning
In Context Semi-Supervised Learning
Jiashuo Fan, Paul Rosu, Aaron T Wang, Lawrence Carin, Xiang Cheng
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 |
|---|---|---|---|
| train end-to-end Transformer with Laplacian-eigenmap-inspired two-stage inductive bias (Orig+E2E-ICL) vs Eig+ICL | ICL Transformer · Cylinder manifold | accuracy: +5-7 percentage points | Improved |
| use end-to-end model instead of strongest baseline | Orig+E2E-ICL · 5-factor product manifold | accuracy: 8-10% margin | Improved |
| train on other manifolds and test OOD on cylinder | Orig+E2E-ICL · Cylinder (OOD) | accuracy: +3-5 percentage points over ID | Improved |
| zero-shot transfer from synthetic manifolds to image manifolds | Orig+E2E-ICL · Stable Diffusion image manifolds | accuracy: 77% at 15% labels; above 62% at 3 labels | Improved |
| use model with built-in structure vs scale-matched baseline Transformer | Orig+E2E-ICL · ImageNet100 (3% labels) | accuracy | Improved |
| increase data for baseline Transformer past ~1200 tasks | Transformer baseline · ImageNet100 (3% labels) | accuracy | Improved |
| replace RBF kernel with linear kernel in ICL head (ablation) | Orig+E2E-ICL · manifold benchmarks | accuracy | No clear effect |
Closest papers by meaning
- Unlabeled Data Can Provably Enhance In-Context Learning of Transformers (NeurIPS 2025 · similarity 0.90)
- How Do Transformers Learn In-Context Beyond Simple Functions? A Case Study on Learning with Representations (ICLR 2024 · similarity 0.90)
- Towards Understanding How Transformers Learn In-context Through a Representation Learning Lens (NeurIPS 2024 · similarity 0.88)
- In-Context Learning with Representations: Contextual Generalization of Trained Transformers (NeurIPS 2024 · similarity 0.88)
- On the Training Convergence of Transformers for In-Context Classification of Gaussian Mixtures (ICML 2025 · similarity 0.88)
- Learning with Partial-Label and Unlabeled Data: A Uniform Treatment for Supervision Redundancy and Insufficiency (ICML 2024 · similarity 0.88)
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.
