A Topological Filter for Learning with Label Noise

Abstract

Noisy labels can impair the performance of deep neural networks. To tackle this problem, in this paper, we propose a new method for filtering label noise. Unlike most existing methods relying on the posterior probability of a noisy classifier, we focus on the much richer spatial behavior of data in the latent representational space. By leveraging the high-order topological information of data, we are able to collect most of the clean data and train a high-quality model. Theoretically we prove that this topological approach is guaranteed to collect the clean data with high probability. Empirical results show that our method outperforms the state-of-the-arts and is robust to a broad spectrum of noise types and levels.

NeurIPS 2020, fixed some typos

Cited by in corpus (8)

• FINE Samples for Learning with Noisy Labels
• Learning Noise Transition Matrix from Only Noisy Labels via Total Variation Regularization
• Localization in the Crowd with Topological Constraints
• Approximating Instance-Dependent Noise via Instance-Confidence Embedding
• NGC: A Unified Framework for Learning with Open-World Noisy Data
• Learning with Noisy Labels by Efficient Transition Matrix Estimation to Combat Label Miscorrection
• Rethinking Noisy Label Models: Labeler-Dependent Noise with Adversarial Awareness
• Friends and Foes in Learning from Noisy Labels