Research map / Detection and segmentation / CNN features
Data Parameters: A New Family of Parameters for Learning a Differentiable Curriculum
Shreyas Saxena, Oncel Tuzel, Dennis DeCoste
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 instance-level parameters to class-level curriculum (ablation) | DCL ResNet18 · ImageNet val | top-1 accuracy (%): 70.8 -> 71.0 | Improved |
| add class-level parameters to instance-level curriculum (ablation) | DCL ResNet18 · WebVision | top-1 accuracy (%): 67.1 -> 67.5 | Improved |
| add data-parameter curriculum (class+instance) | WRN-28-10 · CIFAR100 test | accuracy (%): 80.1 -> 80.8 | Improved |
| add data-parameter curriculum | ResNet18 · ImageNet val | top-1 accuracy (%): 70.3 -> 71.0 | Improved |
| add data-parameter curriculum | ResNet18 · WebVision | top-1 accuracy (%): 66.3 -> 67.5 | Improved |
| add instance-level curriculum only (ablation) | ResNet18 · WebVision | top-1 accuracy (%): 66.3 -> 67.1 | Improved |
| add instance-level data parameters | SSDNet · KITTI car hard | mAP: 78.0 -> 79.3 | Improved |
| add data-parameter curriculum | SSDNet · KITTI car easy | mAP: 92.1 -> 92.8 | Improved |
| add data-parameter curriculum | WRN-28-10 · CIFAR-10 40% label noise | accuracy (%): 67.97 -> 91.10 | Improved |
| add data-parameter curriculum | WRN-28-10 · CIFAR-100 40% label noise | accuracy (%): 50.66 -> 70.93 | Improved |
| replace MentorNet PD weighting with DCL | WRN-28-10 · CIFAR-100 40% label noise | accuracy (%): 56.9 -> 70.93 | Improved |
| add data-parameter curriculum to training with all random labels | VGG16 · CIFAR100 random labels | training accuracy | Improved |
Closest papers by meaning
- On The Power of Curriculum Learning in Training Deep Networks (ICML 2019 · similarity 0.87)
- Boost Self-Supervised Dataset Distillation via Parameterization, Predefined Augmentation, and Approximation (ICLR 2025 · similarity 0.84)
- Hierarchical Features Matter: A Deep Exploration of Progressive Parameterization Method for Dataset Distillation (CVPR 2025 · similarity 0.84)
- Decoupled Optimisation for Long-Tailed Visual Recognition (AAAI 2024 · similarity 0.84)
- SELECT: A Large-Scale Benchmark of Data Curation Strategies for Image Classification (NeurIPS 2024 · similarity 0.84)
- Over-parameterized Student Model via Tensor Decomposition Boosted Knowledge Distillation (NeurIPS 2024 · similarity 0.83)
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.
