Neural Active Learning Meets the Partial Monitoring Framework
This work addresses the challenge of efficient online learning with limited labeled data for machine learning practitioners, though it appears incremental by extending existing frameworks.
The authors tackled the problem of online active learning where agents must balance information acquisition costs with prediction errors by proposing a novel foundation based on partial monitoring theory, and introduced NeuralCBP, which demonstrated favorable performance against state-of-the-art baselines on binary, multi-class, and cost-sensitive tasks.
We focus on the online-based active learning (OAL) setting where an agent operates over a stream of observations and trades-off between the costly acquisition of information (labelled observations) and the cost of prediction errors. We propose a novel foundation for OAL tasks based on partial monitoring, a theoretical framework specialized in online learning from partially informative actions. We show that previously studied binary and multi-class OAL tasks are instances of partial monitoring. We expand the real-world potential of OAL by introducing a new class of cost-sensitive OAL tasks. We propose NeuralCBP, the first PM strategy that accounts for predictive uncertainty with deep neural networks. Our extensive empirical evaluation on open source datasets shows that NeuralCBP has favorable performance against state-of-the-art baselines on multiple binary, multi-class and cost-sensitive OAL tasks.