Learning with Noisy Labels: Interconnection of Two Expectation-Maximizations
This addresses the bottleneck of labor-intensive labeling in computer vision by improving robustness to noisy labels, though it is incremental as it builds on existing LNL methods.
The paper tackles the problem of learning with noisy labels by proposing a framework that models it as finding a structured manifold in noisy data, using two interconnected EM cycles to distinguish clean from corrupted labels and refurbish corrupted ones, achieving state-of-the-art performance with substantial margins on multiple benchmarks.
Labor-intensive labeling becomes a bottleneck in developing computer vision algorithms based on deep learning. For this reason, dealing with imperfect labels has increasingly gained attention and has become an active field of study. We address learning with noisy labels (LNL) problem, which is formalized as a task of finding a structured manifold in the midst of noisy data. In this framework, we provide a proper objective function and an optimization algorithm based on two expectation-maximization (EM) cycles. The separate networks associated with the two EM cycles collaborate to optimize the objective function, where one model is for distinguishing clean labels from corrupted ones while the other is for refurbishing the corrupted labels. This approach results in a non-collapsing LNL-flywheel model in the end. Experiments show that our algorithm achieves state-of-the-art performance in multiple standard benchmarks with substantial margins under various types of label noise.