Decoding class dynamics in learning with noisy labels

Albert Tatjer

doi:10.1016/j.patrec.2024.04.012

Decoding class dynamics in learning with noisy labels

2024-08-01

1

PRE

AI

DOI：10.1016/j.patrec.2024.04.012

原文链接

原文求助

摘要

En 中文

The creation of large-scale datasets annotated by humans inevitably introduces noisy labels, leading to reduced generalization in deep-learning models. Sample selection-based learning with noisy labels is a recent approach that exhibits promising upbeat performance improvements. The selection of clean samples amongst the noisy samples is an important criterion in the learning process of these models. In this work, we delve deeper into the clean-noise split decision and highlight the aspect that effective demarcation of samples would lead to better performance. We identify the Global Noise Conundrum in the existing models, where distribution of samples is treated globally. We propose a per-class-based local distribution of samples and demonstrate the effectiveness of this approach in having a better clean-noise split. We validate our proposal on several benchmarks - both real and synthetic, and show substantial improvements over different stateof-the-art algorithms. We further propose a new metric, classiness to extend our analysis and highlight effectiveness of the proposed method. Source code and instructions to reproduce this paper are available https://github.com/aldakata/CCLM/

Keyword:

Learning with Noisy Labels

Label Noise Modelling

Class Dynamics

Decoding class dynamics in learning with noisy labels

摘要

期刊

机构