Research map / Label-efficient and robust learning / Self-supervised and contrastive
Amortised Invariance Learning for Contrastive Self-Supervision
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| Learn task-specific invariance parameters alongside a linear readout using a shared amortized feature extractor | MoCo ResNet50 · ImageNet-100 downstream suite | Avg. score: 67.8 -> 72.7 | Helped |
| Learn task-specific invariance parameters alongside a linear readout using a shared amortized feature extractor | MoCo ResNet50 · 300w | R²: 85.5 -> 90.0 | Helped |
| Use five invariances rather than two (ablation) | AI-SimCLR ResNet50 · ImageNet-100 downstream suite | Avg. score: 67.1 -> 68.1 | Helped |
| Learn task-specific invariance parameters alongside a linear readout using a shared amortized feature extractor | MoCo-v3 ViT-B · 300w | R²: 81.6 -> 89.0 | Helped |
| Quantize learned invariance parameters (ablation) | AI-MoCo-v3 ViT-B · MoCo-v3 downstream suite | Avg. score: 58.6 -> 56.4 | Hurt |
| Quantize learned invariance parameters (ablation) | AI-MoCo ResNet50 · MoCo downstream suite | Avg. score: 65.7 -> 63.5 | Hurt |
| Learn task-specific invariance parameters alongside a linear readout | SimCLR ResNet18 · VoxCeleb1 (5-way 1-shot) | Accuracy: 36.91 ± 0.19 -> 43.01 ± 0.20 | Helped |
Closest papers by meaning
- Demystifying Contrastive Self-Supervised Learning: Invariances, Augmentations and Dataset Biases (NeurIPS 2020 · similarity 0.89)
- What Should Not Be Contrastive in Contrastive Learning (ICLR 2021 · similarity 0.88)
- On the duality between contrastive and non-contrastive self-supervised learning (ICLR 2023 · similarity 0.88)
- Contrastive-Equivariant Self-Supervised Learning Improves Alignment with Primate Visual Area IT (NeurIPS 2024 · similarity 0.86)
- A Theoretical Analysis of Self-Supervised Learning for Vision Transformers (ICLR 2025 · similarity 0.86)
- Multi-modal contrastive learning adapts to intrinsic dimensions of shared latent variables (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.