Research map / Architectures and efficiency / Pruning and efficient tuning
Pushing the Limits of Simple Pipelines for Few-Shot Learning: External Data and Fine-Tuning Make a Difference
Check 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 BEiT pre-training followed by supervised ImageNet-21k fine-tuning rather than no pre-training | ProtoNet with ViT-base · Meta-Dataset Avg | accuracy: 52.805 -> 82.755 | Helped |
| Increase the probability of support-set data augmentation (ablation) | Fine-tuning phase · Traffic sign and MSCOCO validation domains | performance | Helped |
| Tune the fine-tuning learning rate (ablation) | Fine-tuning phase · Traffic sign and MSCOCO validation domains | performance | Mixed |
| Use 50 gradient steps with an appropriate learning rate (ablation) | Fine-tuning phase · Traffic sign and MSCOCO validation domains | performance | Helped |
| Initialize ProtoNet with pre-trained features rather than training from scratch | ProtoNet · Meta-Dataset | accuracy: 52.805 -> 78.528 | Helped |
Closest papers by meaning
- A Baseline for Few-Shot Image Classification (ICLR 2020 · similarity 0.92)
- Pushing the Limit of Fine-Tuning for Few-Shot Learning: Where Feature Reusing Meets Cross-Scale Attention (AAAI 2024 · similarity 0.91)
- Strong Baselines for Parameter-Efficient Few-Shot Fine-Tuning (AAAI 2024 · similarity 0.91)
- Scaling Few-Shot Learning for the Open World (AAAI 2024 · similarity 0.89)
- One Meta-tuned Transformer is What You Need for Few-shot Learning (ICML 2024 · similarity 0.89)
- Making Large Vision Language Models to Be Good Few-Shot Learners (AAAI 2025 · 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.