Learning to Compare: Relation Network for Few-Shot Learning

CVPR 2018 Flood SungYongxin YangLi ZhangTao XiangPhilip H. S. TorrTimothy M. Hospedales

We present a conceptually simple, flexible, and general framework for few-shot learning, where a classifier must learn to recognise new classes given only few examples from each. Our method, called the Relation Network (RN), is trained end-to-end from scratch... (read more)

PDF Abstract

Evaluation Results from the Paper


TASK DATASET MODEL METRIC NAME METRIC VALUE GLOBAL RANK COMPARE
Few-Shot Image Classification Mini-ImageNet - 1-Shot Learning Relation-Net Accuracy 50.44% # 20
Few-Shot Learning Mini-ImageNet - 1-Shot Learning Relation-Net Accuracy 50.44% # 4
Few-Shot Image Classification Mini-ImageNet - 5-Shot Learning Relation-Net Accuracy 65.32% # 20
Few-Shot Learning Mini-ImageNet - 5-Shot Learning Relation-Net Accuracy 65.32% # 4