Cost-Based Budget Active Learning for Deep Learning
This work aims to improve active learning efficiency for deep learning practitioners by reducing labeling and decision costs, particularly in scenarios where mislabeling is costly.
This paper proposes Cost-Based Budget Active Learning (CBAL) to address the limitations of classical active learning approaches that often select outlier instances and incur high mislabeling costs. CBAL considers classification uncertainty and instance diversity within a budget, using a min-max approach to minimize labeling and decision costs.
Majorly classical Active Learning (AL) approach usually uses statistical theory such as entropy and margin to measure instance utility, however it fails to capture the data distribution information contained in the unlabeled data. This can eventually cause the classifier to select outlier instances to label. Meanwhile, the loss associated with mislabeling an instance in a typical classification task is much higher than the loss associated with the opposite error. To address these challenges, we propose a Cost-Based Bugdet Active Learning (CBAL) which considers the classification uncertainty as well as instance diversity in a population constrained by a budget. A principled approach based on the min-max is considered to minimize both the labeling and decision cost of the selected instances, this ensures a near-optimal results with significantly less computational effort. Extensive experimental results show that the proposed approach outperforms several state-of -the-art active learning approaches.