Research map / Detection and segmentation / Open-set and few-shot detection
Time to augment self-supervised visual representation learning
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 |
|---|---|---|---|
| Combine time-based and conventional augmentations | SimCLR · VHE | category accuracy: 0.328 ± 0.002 -> 0.537 ± 0.001 | Helped |
| Combine time-based and conventional augmentations | SimCLR · 3DShapeE | category accuracy: 0.527 ± 0.005 -> 0.621± 0.001 | Helped |
| Combine time-based and conventional augmentations | SimCLR · ToyBox | category accuracy: 0.313 ± 0.001 -> 0.378 ± 0.005 | Helped |
| Combine time-based and conventional augmentations | SimCLR · CORe50 | category accuracy: 0.544 ± 0.074 -> 0.610 ± 0.090 | Helped |
| Use time-based augmentations alone | SimCLR · ToyBox | category accuracy: 0.313 ± 0.001 -> 0.189 ± 0.009 | Hurt |
| Increase 3-D object rotation speed | SimCLR-TT+ · 3DShapeE | category accuracy | Helped |
| Use 3-D object rotations instead of horizontal flipping (ablation) | SimCLR · 3DShapeE | category accuracy | Helped |
| Add horizontal flip when rotations are already at rot(360) (ablation) | SimCLR · 3DShapeE | category accuracy | No clear effect |
| Move with the same object across changing backgrounds (ablation) | SimCLR-TT+ · VHE, category recognition | category accuracy | Helped |
| Move with the same object across changing backgrounds (ablation) | SimCLR-TT+ · VHE, background classification | accuracy | Hurt |
| Remove object rotations (ablation) | SimCLR-TT+ · ToyBox | category accuracy | Hurt |
| Remove translation-based time augmentations (ablation) | SimCLR-TT+ · ToyBox | category accuracy | No clear effect |
Closest papers by meaning
- Harnessing small projectors and multiple views for efficient vision pretraining (NeurIPS 2024 · similarity 0.86)
- Can Generative Models Improve Self-Supervised Representation Learning? (AAAI 2025 · similarity 0.86)
- No More Shortcuts: Realizing the Potential of Temporal Self-Supervision (AAAI 2024 · similarity 0.85)
- Object Concepts Emerge from Motion (NeurIPS 2025 · similarity 0.85)
- Probing the Mid-level Vision Capabilities of Self-Supervised Learning (CVPR 2025 · similarity 0.85)
- In-Context Symmetries: Self-Supervised Learning through Contextual World Models (NeurIPS 2024 · similarity 0.84)
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.