Active Learning with Imperfect Labels: Optimal Labeler Assignment and Sample Selection

Active learning (AL) is commonly used in applications where labeling data are expensive or time consuming. In practice, however, labels are often noisy because of varying labeler expertise and annotation uncertainty, especially for complex or ambiguous samples. Learning from such imperfectly labeled data can degrade classifier performance. We propose an AL framework that explicitly accounts for label noise by optimally assigning labelers and selecting samples to minimize labeling error. Our approach, which is called optimal labeler assignment and sampling (OLAS), uses a noise model that depends on both labeler accuracy and model uncertainty to guide these decisions. We develop two tractable optimization formulations: one for assigning samples to labelers to minimize worst-case noise and another for selecting samples while controlling overall label noise. Theoretical results provide closed-form solutions under mild conditions. Empirical evaluations on benchmark data sets and a real-world warranty claim classification problem show that OLAS achieves the highest or near-highest classification accuracy among existing AL strategies across most settings using only a single label per sample.