# When Does Self-supervision Improve Few-shot Learning?

arXiv: https://arxiv.org/abs/1910.03560 · HTML: https://mlsanity.com/paper/1910.03560/

## Key facts

- Cited by 78 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 8 Oct 2019.

## Built on by (most cited first, 30 of 78)

- [Benchmarking Representation Learning for Natural World Image Collections](https://mlsanity.com/paper/2103.16483/) (2021)
- [A Comprehensive Survey of Few-shot Learning: Evolution, Applications, Challenges, and Opportunities](https://mlsanity.com/paper/2205.06743/) (2022)
- [Fine-Grained Image Analysis with Deep Learning: A Survey](https://mlsanity.com/paper/2111.06119/) (2021)
- [Simpler is Better: Few-shot Semantic Segmentation with Classifier Weight Transformer](https://mlsanity.com/paper/2108.03032/) (2021)
- [Improving Test-Time Adaptation via Shift-agnostic Weight Regularization and Nearest Source Prototypes](https://mlsanity.com/paper/2207.11707/) (2022)
- [MT3: Meta Test-Time Training for Self-Supervised Test-Time Adaption](https://mlsanity.com/paper/2103.16201/) (2021)
- [Exploring Complementary Strengths of Invariant and Equivariant Representations for Few-Shot Learning](https://mlsanity.com/paper/2103.01315/) (2021)
- [Factors of Influence for Transfer Learning across Diverse Appearance Domains and Task Types](https://mlsanity.com/paper/2103.13318/) (2021)
- [Pareto Self-Supervised Training for Few-Shot Learning](https://mlsanity.com/paper/2104.07841/) (2021)
- [A Realistic Evaluation of Semi-Supervised Learning for Fine-Grained Classification](https://mlsanity.com/paper/2104.00679/) (2021)
- [Rethinking Generalization in Few-Shot Classification](https://mlsanity.com/paper/2206.07267/) (2022)
- [Few-Shot Classification with Contrastive Learning](https://mlsanity.com/paper/2209.08224/) (2022)
- [On Fast Adversarial Robustness Adaptation in Model-Agnostic Meta-Learning](https://mlsanity.com/paper/2102.10454/) (2021)
- [Matching Feature Sets for Few-Shot Image Classification](https://mlsanity.com/paper/2204.00949/) (2022)
- [LibFewShot: A Comprehensive Library for Few-shot Learning](https://mlsanity.com/paper/2109.04898/) (2021)
- [Learning from Future: A Novel Self-Training Framework for Semantic Segmentation](https://mlsanity.com/paper/2209.06993/) (2022)
- [How does unlabeled data improve generalization in self-training? A one-hidden-layer theoretical analysis](https://mlsanity.com/paper/2201.08514/) (2022)
- [Self-Supervision Can Be a Good Few-Shot Learner](https://mlsanity.com/paper/2207.09176/) (2022)
- [Reinforced Attention for Few-Shot Learning and Beyond](https://mlsanity.com/paper/2104.04192/) (2021)
- [Contrastive Prototype Learning with Augmented Embeddings for Few-Shot Learning](https://mlsanity.com/paper/2101.09499/) (2021)
- [Rich Semantics Improve Few-shot Learning](https://mlsanity.com/paper/2104.12709/) (2021)
- [Adaptive Poincar'e Point to Set Distance for Few-Shot Classification](https://mlsanity.com/paper/2112.01719/) (2021)
- [POODLE: Improving Few-shot Learning via Penalizing Out-of-Distribution Samples](https://mlsanity.com/paper/2206.04679/) (2022)
- [On the Importance of Distractors for Few-Shot Classification](https://mlsanity.com/paper/2109.09883/) (2021)
- [Tree Structure-Aware Few-Shot Image Classification via Hierarchical Aggregation](https://mlsanity.com/paper/2207.06989/) (2022)
- [Task Agnostic Representation Consolidation: a Self-supervised based Continual Learning Approach](https://mlsanity.com/paper/2207.06267/) (2022)
- [Multi-level Second-order Few-shot Learning](https://mlsanity.com/paper/2201.05916/) (2022)
- [Semi-Supervised Learning with Taxonomic Labels](https://mlsanity.com/paper/2111.11595/) (2021)
- [A Survey of the Self Supervised Learning Mechanisms for Vision Transformers](https://mlsanity.com/paper/2408.17059/) (2024)
- [Revisiting Local Descriptor for Improved Few-Shot Classification](https://mlsanity.com/paper/2103.16009/) (2021)

## Abstract (arXiv)

We investigate the role of self-supervised learning (SSL) in the context of few-shot learning. Although recent research has shown the benefits of SSL on large unlabeled datasets, its utility on small datasets is relatively unexplored. We find that SSL reduces the relative error rate of few-shot meta-learners by 4%-27%, even when the datasets are small and only utilizing images within the datasets. The improvements are greater when the training set is smaller or the task is more challenging. Although the benefits of SSL may increase with larger training sets, we observe that SSL can hurt the performance when the distributions of images used for meta-learning and SSL are different. We conduct a systematic study by varying the degree of domain shift and analyzing the performance of several meta-learners on a multitude of domains. Based on this analysis we present a technique that automatically selects images for SSL from a large, generic pool of unlabeled images for a given dataset that provides further improvements.