MetaCoCo: A New Few-Shot Classification Benchmark with Spurious Correlation

arXiv:2404.19644v1 Announce Type: new
Abstract: Out-of-distribution (OOD) problems in few-shot classification (FSC) occur when novel classes sampled from testing distributions differ from base classes drawn from training distributions, which considerably degrades the performance of deep learning models deployed in real-world applications. Recent studies suggest that the OOD problems in FSC mainly including: (a) cross-domain few-shot classification (CD-FSC) and (b) spurious-correlation few-shot classification (SC-FSC). Specifically, CD-FSC occurs when a classifier learns transferring knowledge from base classes drawn from seen training distributions …

arxiv benchmark classification correlation cs.cv few-shot type