Research map / Label-efficient and robust learning / Self-supervised and contrastive
Towards Understanding Why Mask Reconstruction Pretraining Helps in Downstream Tasks
Jiachun Pan, Pan Zhou, Shuicheng YAN
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 |
|---|---|---|---|
| Mask-reconstruction pretraining (SimMIM) instead of supervised training on ResNet50 | ResNet50 · ImageNet classification | Top-1 accuracy (%): +1.8 | Improved |
| Mask-reconstruction pretraining instead of supervised training on ResNet50 | ResNet50 · VOC07+12 detection | AP75: +4.4 | Improved |
| Mask-reconstruction pretraining instead of supervised training on ResNet50 | ResNet50 · VOC07 transfer classification | Top-1 accuracy (%): +3.6 | Improved |
| Mask-reconstruction pretraining instead of supervised training on ResNet50 | ResNet50 · ADE20K semantic segmentation | mIoU: +1.9 | Improved |
| Using a teacher network (semantic feature targets) rather than pixel targets in MRP theory analysis | Teacher-Student MRP · theoretical analysis | feature capture guarantee | No clear effect |
Closest papers by meaning
- Understanding and Enhancing Mask-Based Pretraining towards Universal Representations (NeurIPS 2025 · similarity 0.88)
- Video Language Model Pretraining with Spatio-temporal Masking (CVPR 2025 · similarity 0.87)
- Learning What to Predict: Downstream-Guided Task Design for Continued Pretraining (NeurIPS 2026 · similarity 0.87)
- Exploring Target Representations for Masked Autoencoders (ICLR 2024 · similarity 0.85)
- On the Provable Advantage of Unsupervised Pretraining (ICLR 2024 · similarity 0.85)
- Masking meets Supervision: A Strong Learning Alliance (CVPR 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.
