Part-dependent Label Noise: Towards Instance-dependent Label Noise

TL;DR

This paper introduces a novel approach to model and mitigate instance-dependent label noise by decomposing instances into parts and approximating the noise transition matrix through part-based matrices, inspired by human perception.

Contribution

The paper proposes a part-dependent label noise model that approximates instance-dependent noise using part-based transition matrices learned from anchor points.

Findings

Outperforms state-of-the-art methods on synthetic datasets.

Effective in real-world noisy label scenarios.

Demonstrates robustness to complex label noise patterns.

Abstract

Learning with the \textit{instance-dependent} label noise is challenging, because it is hard to model such real-world noise. Note that there are psychological and physiological evidences showing that we humans perceive instances by decomposing them into parts. Annotators are therefore more likely to annotate instances based on the parts rather than the whole instances, where a wrong mapping from parts to classes may cause the instance-dependent label noise. Motivated by this human cognition, in this paper, we approximate the instance-dependent label noise by exploiting \textit{part-dependent} label noise. Specifically, since instances can be approximately reconstructed by a combination of parts, we approximate the instance-dependent \textit{transition matrix} for an instance by a combination of the transition matrices for the parts of the instance. The transition matrices for parts can be learned by exploiting anchor points (i.e., data points that belong to a specific class almost surely). Empirical evaluations on synthetic and real-world datasets demonstrate our method is superior to the state-of-the-art approaches for learning from the instance-dependent label noise.