Non-Convex Boosting Overcomes Random Label Noise
This work addresses the issue of label noise robustness in boosting algorithms for machine learning practitioners, but it is incremental as it compares existing methods rather than introducing new ones.
The study tackled the problem of boosting algorithms' sensitivity to random label noise by evaluating AdaBoost, LogitBoost, BrownBoost, and RobustBoost on synthetic and real datasets with corrupted labels. It found that BrownBoost and RobustBoost performed significantly better than AdaBoost and LogitBoost in the presence of noise, with no significant differences between each pair of algorithms, and provided an explanation based on margin distributions.
The sensitivity of Adaboost to random label noise is a well-studied problem. LogitBoost, BrownBoost and RobustBoost are boosting algorithms claimed to be less sensitive to noise than AdaBoost. We present the results of experiments evaluating these algorithms on both synthetic and real datasets. We compare the performance on each of datasets when the labels are corrupted by different levels of independent label noise. In presence of random label noise, we found that BrownBoost and RobustBoost perform significantly better than AdaBoost and LogitBoost, while the difference between each pair of algorithms is insignificant. We provide an explanation for the difference based on the margin distributions of the algorithms.