Abstract:Deep active learning (DeepAL) offers a viable solution for enhancing the predictive performance of models with limited annotation costs. Popular DeepAL pipelines fail to (1) adequately and simultaneously draw on unlabeled and labeled samples to learn accurate data representations, and (2) progressively balance the distribution of selected samples across classes during the query. In this paper, we introduce a self-supervised DeepAL (SSAL) framework as a solution to the aforementioned challenges. First, SSAL simultaneously utilizes both labeled data and unlabeled data to learn high-quality representations. The method learns the image rotation degree for unlabeled data in a self-supervised manner, which can be jointly processed with the supervised learning. Second, the sample information is interpreted by fusing the discriminant uncertainty and the presumed mastery ability of each sample. This approach prevents the selection of many nonrepresentative boundary samples based solely on uncertainty and avoids the selection of aggregated samples based solely on clustering. Third, we propose a new sample selection strategy named class rebalanced group sampling. This strategy rebalances classes by the optimal size of query samples for each class, which is induced from the approximation between the population and the model distribution. Experiments were conducted on five public image datasets, namely CIFAR-10/100, FashionMNIST, SVHN, and TinyImageNet, and the results demonstrated the effectiveness of our approach. The code is available at https://github.com/FanSmale/SSAL.

Learning Generalizable Batch Active Learning Strategies Via Deep Q-networks (student Abstract).

TBAL: Two-stage batch-mode active learning for image classification

GraphCBAL: Class-Balanced Active Learning for Graph Neural Networks Via Reinforcement Learning

Bidirectional Uncertainty-Based Active Learning for Open Set Annotation

Meta-Learning Transferable Active Learning Policies by Deep Reinforcement Learning

Deep active learning models for imbalanced image classification

Distributed Active Learning.

A Multicriterion Query-Based Batch Mode Active Learning Technique

Self-supervised Class-Balanced Active Learning with Uncertainty-Mastery Fusion

Bio-inspired Active Learning Method in Spiking Neural Network.

New Balanced Active Learning Model and Optimization Algorithm.

An Empirical Study on the Efficacy of Deep Active Learning Techniques

Discovering General-Purpose Active Learning Strategies

Batch Active Learning with Graph Neural Networks via Multi-Agent Deep Reinforcement Learning

Active Learning Methods with Deep Gaussian Processes.

Frugal Reinforcement-based Active Learning

GALAXY: Graph-based Active Learning at the Extreme

Cost-effective Batch-mode Multi-label Active Learning

BALQUE: Batch active learning by querying unstable examples with calibrated confidence

Loss-based Active Learning Via Double-Branch Deep Network

Image Classification with Deep Reinforcement Active Learning