Research map / Label-efficient and robust learning / Zero-/few-shot and meta-learning
Meta-Learning with Fewer Tasks through Task Interpolation
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 |
|---|---|---|---|
| Add task interpolation | MAML · Pose (15-shot) | MSE: 2.383 -> 1.976 | Helped |
| Add task interpolation | MAML · RMNIST (1-shot) | accuracy: 57.34 -> 65.92 | Helped |
| Add task interpolation | MAML · NCI (5-shot) | accuracy: 77.09 -> 79.14 | Helped |
| Add task interpolation | MAML · Metabolism (5-shot) | accuracy: 57.22 -> 60.28 | Helped |
| Add task interpolation | MAML · miniImagenet-S (1-shot) | accuracy: 38.27% -> 41.58% | Helped |
| Add task interpolation | MAML · ISIC (5-shot) | accuracy: 65.24% -> 70.69% | Helped |
| Add intra-task interpolation to the full task-interpolation method (ablation) | MAML · RMNIST (1-shot) | accuracy: 65.34 -> 65.92 | Helped |
| Add cross-task interpolation to the full task-interpolation method (ablation) | MAML · RMNIST (1-shot) | accuracy: 62.57 -> 65.92 | Helped |
| Add intra-task interpolation to the full task-interpolation method (ablation) | ProtoNet · RMNIST (1-shot) | accuracy: 69.97 -> 70.14 | Helped |
| Add cross-task interpolation to the full task-interpolation method (ablation) | ProtoNet · RMNIST (1-shot) | accuracy: 67.32 -> 70.14 | Helped |
Closest papers by meaning
- Set-based Meta-Interpolation for Few-Task Meta-Learning (NeurIPS 2022 · similarity 0.92)
- Meta-Learning without Memorization (ICLR 2020 · similarity 0.90)
- Bridging Multi-Task Learning and Meta-Learning: Towards Efficient Training and Effective Adaptation (ICML 2021 · similarity 0.88)
- Provable Meta-Learning of Linear Representations (ICML 2021 · similarity 0.88)
- How to Distribute Data across Tasks for Meta-Learning? (AAAI 2022 · similarity 0.88)
- Unsupervised Meta-Learning via In-Context Learning (ICLR 2025 · similarity 0.87)
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.