Salvage Reusable Samples from Noisy Data for Robust Learning

TL;DR

This paper introduces CRSSC, a method that identifies, corrects, and reuses noisy samples in web image data to improve the robustness of deep fine-grained models, outperforming existing techniques.

Contribution

The paper proposes a novel certainty-based sample selection and correction method that reuses noisy samples, enhancing deep FG model training with web images.

Findings

CRSSC outperforms existing noise-handling methods in experiments.

Theoretical analysis confirms the effectiveness of sample correction.

Reusing samples boosts model robustness and accuracy.

Abstract

Due to the existence of label noise in web images and the high memorization capacity of deep neural networks, training deep fine-grained (FG) models directly through web images tends to have an inferior recognition ability. In the literature, to alleviate this issue, loss correction methods try to estimate the noise transition matrix, but the inevitable false correction would cause severe accumulated errors. Sample selection methods identify clean ("easy") samples based on the fact that small losses can alleviate the accumulated errors. However, "hard" and mislabeled examples that can both boost the robustness of FG models are also dropped. To this end, we propose a certainty-based reusable sample selection and correction approach, termed as CRSSC, for coping with label noise in training deep FG models with web images. Our key idea is to additionally identify and correct reusable samples, and then leverage them together with clean examples to update the networks. We demonstrate the superiority of the proposed approach from both theoretical and experimental perspectives.