# Selecting Relevant Features from a Multi-domain Representation for Few-shot Classification

arXiv: https://arxiv.org/abs/2003.09338 · HTML: https://mlsanity.com/paper/2003.09338/

## Key facts

- Cited by 11 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 20 Mar 2020.

## Built on by (most cited first, 11 of 11)

- [FedProto: Federated Prototype Learning across Heterogeneous Clients](https://mlsanity.com/paper/2105.00243/) (2021)
- [Orthogonal Projection Loss](https://mlsanity.com/paper/2103.14021/) (2021)
- [A linearized framework and a new benchmark for model selection for fine-tuning](https://mlsanity.com/paper/2102.00084/) (2021)
- [Meta-Album: Multi-domain Meta-Dataset for Few-Shot Image Classification](https://mlsanity.com/paper/2302.08909/) (2023)
- [Impact of Aliasing on Generalization in Deep Convolutional Networks](https://mlsanity.com/paper/2108.03489/) (2021)
- [Anti-aliasing Semantic Reconstruction for Few-Shot Semantic Segmentation](https://mlsanity.com/paper/2106.00184/) (2021)
- [Supervised Momentum Contrastive Learning for Few-Shot Classification](https://mlsanity.com/paper/2101.11058/) (2021)
- [Meta Omnium: A Benchmark for General-Purpose Learning-to-Learn](https://mlsanity.com/paper/2305.07625/) (2023)
- [Neural Fine-Tuning Search for Few-Shot Learning](https://mlsanity.com/paper/2306.09295/) (2023)
- [NeurIPS'22 Cross-Domain MetaDL competition: Design and baseline results](https://mlsanity.com/paper/2208.14686/) (2022)
- [Out-of-distribution Few-shot Learning For Edge Devices without Model Fine-tuning](https://mlsanity.com/paper/2304.06309/) (2023)

## Abstract (arXiv)

Popular approaches for few-shot classification consist of first learning a generic data representation based on a large annotated dataset, before adapting the representation to new classes given only a few labeled samples. In this work, we propose a new strategy based on feature selection, which is both simpler and more effective than previous feature adaptation approaches. First, we obtain a multi-domain representation by training a set of semantically different feature extractors. Then, given a few-shot learning task, we use our multi-domain feature bank to automatically select the most relevant representations. We show that a simple non-parametric classifier built on top of such features produces high accuracy and generalizes to domains never seen during training, which leads to state-of-the-art results on MetaDataset and improved accuracy on mini-ImageNet.