Dense Classification and Implanting for Few-Shot Learning

Abstract : Training deep neural networks from few examples is a highly challenging and key problem for many computer vision tasks. In this context, we are targeting knowledge transfer from a set with abundant data to other sets with few available examples. We propose two simple and effective solutions: (i) dense classification over feature maps, which for the first time studies local activations in the domain of few-shot learning, and (ii) implanting, that is, attaching new neurons to a previously trained network to learn new, task-specific features. On miniImageNet, we improve the prior state-of-the-art on few-shot classification, i.e., we achieve 62.5%, 79.8% and 83.8% on 5-way 1-shot, 5-shot and 10-shot settings respectively.
Document type :
Preprints, Working Papers, ...
Complete list of metadatas

https://hal.inria.fr/hal-02370192
Contributor : Yannis Avrithis <>
Submitted on : Tuesday, November 19, 2019 - 12:44:58 PM
Last modification on : Saturday, November 23, 2019 - 1:37:06 AM

File

1903.05050.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : hal-02370192, version 1
  • ARXIV : 1903.05050

Citation

Yann Lifchitz, Yannis Avrithis, Sylvaine Picard, Andrei Bursuc. Dense Classification and Implanting for Few-Shot Learning. 2019. ⟨hal-02370192⟩

Share

Metrics

Record views

12

Files downloads

54