Research map / Label-efficient and robust learning / Self-supervised and contrastive
Multi-Task Self-Supervised Visual Learning
Carl Doersch, Andrew Zisserman
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 colorization to relative position | RP · ImageNet frozen (top-5 recall) | recall@5: 59.21 -> 66.64 | Improved |
| add exemplar to relative position | RP · ImageNet frozen (top-5 recall) | recall@5: 59.21 -> 65.24 | Improved |
| add motion segmentation to relative position | RP · ImageNet frozen (top-5 recall) | recall@5: 59.21 -> 63.73 | Improved |
| add colorization, exemplar and motion segmentation to relative position | RP · PASCAL VOC 2007 detection | mAP: 66.75 -> 70.53 | Improved |
| add exemplar to relative position | RP · NYU v2 depth | % pixels below 1.25: 80.54 -> 78.70 | Hurt |
| add input harmonization (grayscale for relative position) (ablation) | RP+Col · ImageNet frozen (top-5 recall) | recall@5: 66.64 -> 68.08 | Improved |
| add input harmonization (grayscale for relative position) (ablation) | RP+Col · PASCAL VOC 2007 detection | mAP: 68.75 -> 68.26 | No clear effect |
| add input harmonization (grayscale for relative position) (ablation) | RP · ImageNet frozen (top-5 recall) | recall@5: 59.21 -> 62.33 | Improved |
| add lasso at evaluation only (ablation) | RP+Col+Ex+MS · ImageNet frozen (top-5 recall) | recall@5: 69.30 -> 70.18 | Improved |
| add lasso at evaluation only (ablation) | RP+Col+Ex+MS · PASCAL VOC 2007 detection | mAP: 70.53 -> 68.86 | Hurt |
| add lasso at pre-training only (ablation) | RP+Col+Ex+MS · ImageNet frozen (top-5 recall) | recall@5: 69.30 -> 68.09 | Hurt |
| add lasso at pre-training and evaluation (ablation) | RP+Col+Ex+MS · ImageNet frozen (top-5 recall) | recall@5: 69.30 -> 69.44 | No clear effect |
Closest papers by meaning
- Multi-Task Self-Training for Learning General Representations (ICCV 2021 · similarity 0.89)
- Unsupervised Representation Transfer for Small Networks: I Believe I Can Distill On-the-Fly (NeurIPS 2021 · similarity 0.89)
- Integrated perception with recurrent multi-task neural networks (NeurIPS 2016 · similarity 0.89)
- Revisiting Self-Supervised Visual Representation Learning (CVPR 2019 · similarity 0.88)
- Universal Representations: A Unified Look at Multiple Task and Domain Learning (IJCV 2023 · similarity 0.88)
- Self-Supervised Visual Feature Learning With Deep Neural Networks: A Survey (TPAMI 2020 · 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.
