Research map / Label-efficient and robust learning / Zero-/few-shot and meta-learning
‘Less Than One’-Shot Learning: Learning N Classes From M < N Samples
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 probabilistic soft labels on two prototypes | SLaPkNN with k=2 · Euclidean space | classes separable: two soft-label prototypes; n ∈ [1,∞) classes | Helped |
| Use soft labels rather than hard labels for prototype generation | kNN on concentric circles · Concentric circles | prototype requirement: from O(N^2) down to O(1) | Helped |
| Use five soft-label prototypes with SLaPkNN instead of the hard-label 1NN configuration | Circle classification · Concentric circles | prototypes required to separate six circles: five soft-label prototypes | Helped |
| Increase k from 1 to 2 (ablation) | SLaPkNN decision landscapes · LO-shot configurations | decision landscape | Mixed |
| Increase k from 2 to 3 (ablation) | SLaPkNN decision landscapes · LO-shot configurations | risk at decision boundaries | Hurt |
| Tune prototype positions and labels to control decision boundaries and risk | SLaPkNN · LO-shot configurations | robustness to prototype perturbations | Helped |
| Use unrestricted soft labels and vary k | SLaPkNN · Decision landscapes | decision-boundary configurations | Helped |
Closest papers by meaning
- Dense Classification and Implanting for Few-Shot Learning (CVPR 2019 · similarity 0.89)
- A Baseline for Few-Shot Image Classification (ICLR 2020 · similarity 0.89)
- Matching Networks for One Shot Learning (NeurIPS 2016 · similarity 0.89)
- Efficient Training of Language Models using Few-Shot Learning (ICML 2023 · similarity 0.89)
- Learning feed-forward one-shot learners (NeurIPS 2016 · similarity 0.88)
- Few-Shot Learning via Embedding Adaptation With Set-to-Set Functions (CVPR 2020 · 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.