Research map / Label-efficient and robust learning / Knowledge distillation
Collaborative Group Learning
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 |
|---|---|---|---|
| Use Collaborative Group Learning instead of OKDDip | Collaborative learning · CIFAR-10 | Top-1 accuracy (%): 94.29 ± 0.04 -> 94.61 ± 0.06 | Helped |
| Use Collaborative Group Learning instead of OKDDip | Collaborative learning · CIFAR-100 | Top-1 accuracy (%): 76.69 ± 0.04 -> 78.01 ± 0.07 | Helped |
| Use Collaborative Group Learning instead of OKDDip | Collaborative learning · CIFAR-100 | Top-1 accuracy (%): 77.39 ± 0.09 -> 78.31 ± 0.10 | Helped |
| Use Collaborative Group Learning instead of OKDDip | Collaborative learning · Yelp Review Full | Top-1 accuracy (%): 62.75 ± 0.18 -> 63.32 ± 0.04 | Helped |
| Use Collaborative Group Learning instead of OKDDip | Collaborative learning · Yahoo! Answers | Top-1 accuracy (%): 70.10 ± 0.09 -> 70.35 ± 0.05 | Helped |
| Use Collaborative Group Learning instead of OKDDip | Collaborative learning · Amazon Review Full | Top-1 accuracy (%): 60.63 ± 0.04 -> 61.03 ± 0.05 | Helped |
| Use Collaborative Group Learning instead of CL-ILR | Collaborative learning · IMDB Review | Top-1 accuracy (%): 83.10 ± 0.07 -> 83.81 ± 0.10 | Helped |
| Add random routing and module parameter sharing (ablation) | CGL · CIFAR-100 | Top-1 accuracy (%): 75.75 -> 78.01 | Helped |
| Add sub-set data learning (ablation) | CGL · CIFAR-100 | Top-1 accuracy (%): 77.48 -> 78.01 | Helped |
| Add sub-group imitation rather than imitating all other students (ablation) | CGL · CIFAR-100 | Top-1 accuracy (%): 77.23 -> 78.01 | Helped |
| Use a 20% ramp-up range (ablation) | CGL · CIFAR-100 | Top-1 accuracy (%): 77.79 -> 78.01 | Helped |
| Extend ramp-up range from 20% to 40% (ablation) | CGL · CIFAR-100 | Top-1 accuracy (%): 78.01 -> 77.22 | Hurt |
Closest papers by meaning
- Learning Group Actions on Latent Representations (NeurIPS 2024 · similarity 0.85)
- Learning from Teaching Regularization: Generalizable Correlations Should be Easy to Imitate (NeurIPS 2024 · similarity 0.85)
- Mutual Learning for SAM Adaptation: A Dual Collaborative Network Framework for Source-Free Domain Transfer (ICML 2025 · similarity 0.85)
- A Kernel Perspective on Distillation-based Collaborative Learning (NeurIPS 2024 · similarity 0.84)
- Task Groupings Regularization: Data-Free Meta-Learning with Heterogeneous Pre-trained Models (ICML 2024 · similarity 0.84)
- Boosting Residual Networks with Group Knowledge (AAAI 2024 · similarity 0.84)
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.