# On the Importance of Distractors for Few-Shot Classification

Source: [On the Importance of Distractors for Few-Shot Classification](<https://www.sota2.com/research/paper/68fa6acf1e209131cea2a6b0>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2025-10-25T23:33:16.767Z

Authors: Rajshekhar Das, Yu-Xiong Wang, Jos\\'eM.F. Moura

Published: 2021-09-20T23:35:56.000Z

## Abstract

Few-shot classification aims at classifying categories of a novel task by learning from just a few (typically, 1 to 5) labelled examples. An effective approach to few-shot classification involves a prior model trained on a large-sample base domain, which is then finetuned over the novel few-shot task to yield generalizable representations. However, task-specific finetuning is prone to overfitting due to the lack of enough training examples. To alleviate this issue, we propose a new finetuning approach based on contrastive learning that reuses unlabelled examples from the base domain in the form of distractors. Unlike the nature of unlabelled data used in prior works, distractors belong to classes that do not overlap with the novel categories. We demonstrate for the first time that inclusion of such distractors can significantly boost few-shot generalization. Our technical novelty includes a stochastic pairing of examples sharing the same category in the few-shot task and a weighting term that controls the relative influence of task-specific negatives and distractors. An important aspect of our finetuning objective is that it is agnostic to distractor labels and hence applicable to various base domain settings. Compared to state-of-the-art approaches, our method shows accuracy gains of up to $12\\%$ in cross-domain and up to $5\\%$ in unsupervised prior-learning settings.

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

- [Image Classification on BSCD-FSL 8 Domains (test)](<https://www.sota2.com/research/sota/image-classification-on-bscd-fsl-8-domains-test.md>): Image Classification; BSCD-FSL 8 target domains: CUB, Cars, Plantae, Places, CropDisease, EuroSAT, ISIC, ChestX standard (test); CUB Accuracy: 70.53.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/2109.09883>)

[PDF](<https://arxiv.org/pdf/2109.09883.pdf>)