Research map / Label-efficient and robust learning / Zero-/few-shot and meta-learning
Few-Shot Learning Through an Information Retrieval Lens
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 with an mAP objective that jointly uses all within-batch query rankings rather than pairwise Siamese training | Few-shot metric learning · mini-ImageNet, few-shot classification and retrieval | accuracy and mAP | Helped |
| Use the mAP structured objective instead of all-pairs Siamese training when the number of examples per class is small | Few-shot metric learning · mini-ImageNet, low-shot training batches | accuracy and mAP | Helped |
| Optimize the task loss directly with mAP-DLM rather than using the mAP-SSVM upper-bound objective | mAP few-shot learning · Omniglot and mini-ImageNet | accuracy and mAP | No clear effect |
| Use mAP-DLM instead of Siamese training | Few-shot metric learning · Omniglot, 1-shot 20-way retrieval | mAP: 95.7 -> 95.8 | Helped |
| Use mAP-SSVM instead of Siamese training | Few-shot metric learning · mini-ImageNet, 1-shot 5-way retrieval | mAP: 22.66 ± 0.13 % -> 23.87 ± 0.14 % | Helped |
Closest papers by meaning
- Few-Shot Recognition via Stage-Wise Retrieval-Augmented Finetuning (CVPR 2025 · similarity 0.89)
- Finding Task-Relevant Features for Few-Shot Learning by Category Traversal (CVPR 2019 · similarity 0.88)
- Towards Practical Few-shot Query Sets: Transductive Minimum Description Length Inference (NeurIPS 2022 · similarity 0.87)
- Information Maximization for Few-Shot Learning (NeurIPS 2020 · similarity 0.87)
- RankDNN: Learning to Rank for Few-Shot Learning (AAAI 2023 · similarity 0.87)
- Scaling Few-Shot Learning for the Open World (AAAI 2024 · similarity 0.86)
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.