Research map / Label-efficient and robust learning / Self-supervised and contrastive
DATA: Domain-Aware and Task-Aware Self-Supervised Learning
Check 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 |
|---|---|---|---|
| train and select weight-sharing subnets with DATA | MoCo v2 · ImageNet-1% semi-supervised classification | top-1 accuracy: 39.8 -> 49.5 | Helped |
| train and select weight-sharing subnets with DATA | MoCo v2 · PASCAL VOC object detection | AP: 57.2 -> 58.5 | Helped |
| combine DATA with the base SSL method | BYOL · ImageNet-1% classification | top-1 accuracy: 23.7 -> 24.8 | Helped |
| combine DATA with the base SSL method | ReSSL · ImageNet-1% classification | top-1 accuracy: 23.7 -> 26.7 | Helped |
| combine DATA with the base SSL method | DenseCL · PASCAL VOC object detection | AP: 49.1 -> 50.2 | Helped |
| fix the teacher branch to the largest architecture (ablation) | DATA supernet training · ImageNet-10% linear evaluation | top-1 accuracy: 39.0% -> 42.4% | Helped |
| use task-matched global feature z for model selection (ablation) | DATA on ImageNet-1% classification · ImageNet-1% semi-supervised classification | top-1 accuracy: 47.1 -> 47.5 | Helped |
| use task-matched C5 features for model selection (ablation) | DATA on VOC · VOC object detection | AP: 57.8 -> 58.5 | Helped |
| use task-matched C2-C5 features for model selection (ablation) | DATA on COCO · COCO object detection | AP@IoU: 39.2 -> 39.9 | Helped |
| search on the target domain rather than ImageNet (ablation) | DATA on Cityscapes segmentation · Cityscapes semantic segmentation | mIoU: 76.2 -> 77.4 | Helped |
| search on the target domain rather than ImageNet (ablation) | DATA on Cityscapes segmentation · Cityscapes semantic segmentation | correlation: 0.23 -> 0.63 | Helped |
Closest papers by meaning
- Effective Self-supervised Pre-training on Low-compute Networks without Distillation (ICLR 2023 · similarity 0.90)
- TransNAS-Bench-101: Improving Transferability and Generalizability of Cross-Task Neural Architecture Search (CVPR 2021 · similarity 0.89)
- Task-Customized Self-Supervised Pre-training with Scalable Dynamic Routing (AAAI 2022 · similarity 0.88)
- Training-free Neural Architecture Search through Variance of Knowledge of Deep Network Weights (CVPR 2025 · similarity 0.86)
- Subnet-Aware Dynamic Supernet Training for Neural Architecture Search (CVPR 2025 · similarity 0.86)
- Efficient Few-Shot Neural Architecture Search by Counting the Number of Nonlinear Functions (AAAI 2025 · 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.