Research map / Detection and segmentation / Open-set and few-shot detection
Unsupervised Representation Learning by Predicting Image Rotations
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 |
|---|---|---|---|
| Train the pretext task with 4 rotations instead of 8 (ablation) | RotNet · CIFAR-10 classification | accuracy: 88.51 -> 89.06 | Helped |
| Train the pretext task with 4 rotations instead of 2 rotations (0° and 180°) (ablation) | RotNet · CIFAR-10 classification | accuracy: 87.46 -> 89.06 | Helped |
| Train the pretext task with 4 rotations instead of 2 rotations (90° and 270°) (ablation) | RotNet · CIFAR-10 classification | accuracy: 85.52 -> 89.06 | Helped |
| Use 0° and 180° rotations rather than 90° and 270° rotations (ablation) | 2-rotation RotNet · CIFAR-10 classification | accuracy: 85.52 -> 87.46 | Helped |
| Increase RotNet depth from 3 to 4 convolutional blocks (ablation) | RotNet · CIFAR-10 classification, ConvB2 features | accuracy: 88.26 -> 89.06 | Helped |
| Increase RotNet depth from 4 to 5 convolutional blocks (ablation) | RotNet · CIFAR-10 classification, ConvB2 features | accuracy: 89.06 -> 89.76 | Helped |
| Fine-tune the RotNet features during object recognition training (ablation) | RotNet + conv · CIFAR-10 | accuracy: 91.16 -> 92.17 | Helped |
Closest papers by meaning
- Beyond Supervised vs. Unsupervised: Representative Benchmarking and Analysis of Image Representation Learning (CVPR 2022 · similarity 0.90)
- Self-Supervised Representation Learning by Rotation Feature Decoupling (CVPR 2019 · similarity 0.89)
- Unsupervised Learning of Discriminative Attributes and Visual Representations (CVPR 2016 · similarity 0.88)
- Unsupervised Deep Learning by Neighbourhood Discovery (ICML 2019 · similarity 0.88)
- Train a One-Million-Way Instance Classifier for Unsupervised Visual Representation Learning (AAAI 2021 · similarity 0.88)
- Spatially Invariant Unsupervised Object Detection with Convolutional Neural Networks (AAAI 2019 · 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.