Graph Convolutional Networks for Learning with Few Clean and many Noisy Labels

Ahmet Iscen 1 Giorgos Tolias 2 Yannis Avrithis 1 Ondřej Chum 2 Cordelia Schmid 3
1 LinkMedia - Creating and exploiting explicit links between multimedia fragments
Inria Rennes – Bretagne Atlantique , IRISA-D6 - MEDIA ET INTERACTIONS
2 VRG - Visual Recognition Group [Prague]
CTU/FEE - Faculty of electrical engineering [Prague]
3 Thoth - Apprentissage de modèles à partir de données massives
Inria Grenoble - Rhône-Alpes, LJK - Laboratoire Jean Kuntzmann
Abstract : In this work we consider the problem of learning a classifier from noisy labels when a few clean labeled examples are given. The structure of clean and noisy data is modeled by a graph per class and Graph Convolutional Networks (GCN) are used to predict class relevance of noisy examples. For each class, the GCN is treated as a binary classifier learning to discriminate clean from noisy examples using a weighted binary cross-entropy loss function, and then the GCN-inferred "clean" probability is exploited as a relevance measure. Each noisy example is weighted by its relevance when learning a classifier for the end task. We evaluate our method on an extended version of a few-shot learning problem, where the few clean examples of novel classes are supplemented with additional noisy data. Experimental results show that our GCN-based cleaning process significantly improves the classification accuracy over not cleaning the noisy data and standard few-shot classification where only few clean examples are used. The proposed GCN-based method outperforms the transductive approach (Douze et al., 2018) that is using the same additional data without labels.
Document type :
Preprints, Working Papers, ...
Complete list of metadatas

Cited literature [45 references]  Display  Hide  Download

https://hal.inria.fr/hal-02370212
Contributor : Yannis Avrithis <>
Submitted on : Tuesday, November 19, 2019 - 12:54:42 PM
Last modification on : Thursday, November 28, 2019 - 11:59:00 AM

File

1910.00324.pdf
Files produced by the author(s)

Identifiers

Citation

Ahmet Iscen, Giorgos Tolias, Yannis Avrithis, Ondřej Chum, Cordelia Schmid. Graph Convolutional Networks for Learning with Few Clean and many Noisy Labels. 2019. ⟨hal-02370212⟩

Share

Metrics

Record views

29

Files downloads

124