Research map / Label-efficient and robust learning / Self-supervised and contrastive
The SSL Interplay: Augmentations, Inductive Bias, and Generalization
Vivien Cabannes, Bobak Kiani, Randall Balestriero, Yann LeCun, Alberto Bietti
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 regularization lambda (lambda K^{-1} bias, i.e. T_lambda) to the spectral pretraining loss (ablation) | VICReg-style population loss with kernel/architecture inductive bias · Theory: downstream excess risk on sphere S^7 polynomial targets | downstream empirical risk (qualitative) | Mixed |
| Use multiple augmentations per sample (multi-crop) versus more fresh inputs when total m x n is fixed | Pretraining with VICReg-style loss · Theory: pretraining convergence | pretraining convergence rate | No clear effect |
| Negligible regularization (no inductive bias) in empirical pretraining with few augmentations (ablation) | Empirical estimate of T_lambda · Gaussian kernel illustration (capacity trouble) | recovery of cluster structure (qualitative) | Hurt |
| Use stronger augmentations (reduce positive eigenspace of T_lambda) to avoid collapse | SSL pretraining guidance · Practical guideline (theory-based, not experimentally verified) | collapse (qualitative) | No clear effect |
Closest papers by meaning
- Memorization in Self-Supervised Learning Improves Downstream Generalization (ICLR 2024 · similarity 0.89)
- Improving Self-Supervised Learning by Characterizing Idealized Representations (NeurIPS 2022 · similarity 0.89)
- A Closer Look at Benchmarking Self-supervised Pre-training with Image Classification (IJCV 2025 · similarity 0.89)
- Self-supervised Representation Learning from Random Data Projectors (ICLR 2024 · similarity 0.88)
- Views Can Be Deceiving: Improved SSL Through Feature Space Augmentation (ICLR 2024 · similarity 0.88)
- Harnessing small projectors and multiple views for efficient vision pretraining (NeurIPS 2024 · similarity 0.87)
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.
