Squeezing Backbone Feature Distributions to the Max for Efficient Few-Shot Learning

18 Oct 2021  ·  Yuqing Hu, Vincent Gripon, Stéphane Pateux ·

Few-shot classification is a challenging problem due to the uncertainty caused by using few labelled samples. In the past few years, many methods have been proposed with the common aim of transferring knowledge acquired on a previously solved task, what is often achieved by using a pretrained feature extractor... Following this vein, in this paper we propose a novel transfer-based method which aims at processing the feature vectors so that they become closer to Gaussian-like distributions, resulting in increased accuracy. In the case of transductive few-shot learning where unlabelled test samples are available during training, we also introduce an optimal-transport inspired algorithm to boost even further the achieved performance. Using standardized vision benchmarks, we show the ability of the proposed methodology to achieve state-of-the-art accuracy with various datasets, backbone architectures and few-shot settings. read more

PDF Abstract
Task Dataset Model Metric Name Metric Value Global Rank Uses Extra
Training Data
Benchmark
Few-Shot Image Classification CIFAR-FS 5-way (1-shot) PEMnE-BMS* Accuracy 88.44 # 1
Few-Shot Image Classification CIFAR-FS 5-way (5-shot) PEMnE-BMS* Accuracy 91.86 # 1
Few-Shot Image Classification CUB 200 5-way 1-shot PEMnE-BMS* Accuracy 94.78 # 2
Few-Shot Image Classification CUB 200 5-way 5-shot PEMnE-BMS* Accuracy 96.43 # 1
Few-Shot Image Classification Mini-Imagenet 5-way (1-shot) PEMbE-NCM (inductive) Accuracy 68.43 # 22
Few-Shot Image Classification Mini-Imagenet 5-way (1-shot) PEMnE-BMS* (transductive) Accuracy 85.54 # 1
Few-Shot Image Classification Mini-Imagenet 5-way (5-shot) PEMnE-BMS*(transductive) Accuracy 91.53 # 1
Few-Shot Image Classification Mini-Imagenet 5-way (5-shot) PEMbE-NCM (inductive) Accuracy 84.67 # 11
Few-Shot Image Classification Mini-ImageNet-CUB 5-way (1-shot) PEMnE-BMS* Accuracy 63.90 # 1
Few-Shot Image Classification Mini-ImageNet-CUB 5-way (5-shot) PEMnE-BMS Accuracy 79.15 # 1
Few-Shot Image Classification Tiered ImageNet 5-way (1-shot) PEMnE-BMS* Accuracy 86.07 # 1
Few-Shot Image Classification Tiered ImageNet 5-way (5-shot) PEMnE-BMS* Accuracy 91.09 # 1

Methods


No methods listed for this paper. Add relevant methods here