0

Benchmarking Low-Shot Robustness to Natural Distribution Shifts

A study examines robustness to natural distribution shifts in low-shot learning, revealing that no single model or intervention consistently outperforms others across various datasets and architectural settings.

Year
2023
Venue
ICCV 2023 1
Authors
4
Hosting
Abstract onlyARXIV-DEFAULT

Cite

Notes

Only stored in your browser.

Attribution

Abstract & full text
arxiv.org/abs/2304.11263v2ARXIV-DEFAULT
TL;DR
Semantic Scholar
Attribution policy →

Abstract

Robustness to natural distribution shifts has seen remarkable progress thanks to recent pre-training strategies combined with better fine-tuning methods. However, such fine-tuning assumes access to large amounts of labelled data, and the extent to which the observations hold when the amount of training data is not as high remains unknown. We address this gap by performing the first in-depth study of robustness to various natural distribution shifts in different low-shot regimes: spanning datasets, architectures, pre-trained initializations, and state-of-the-art robustness interventions. Most importantly, we find that there is no single model of choice that is often more robust than others, and existing interventions can fail to improve robustness on some datasets even if they do so in the full-shot regime. We hope that our work will motivate the community to focus on this problem of practical importance.

Authors

4