Active Learning is Planning: Nonmyopic ϵ-Bayes-Optimal Active Learning of Gaussian Processes
Published in Machine Learning and Knowledge Discovery in Databases - European Conference, ECML/PKDD-14 NECTAR (New Scientific and Technical Advances in Research), 2014
Abstract: A fundamental issue in active learning of Gaussian processes is that of the exploration-exploitation trade-off. This paper presents a novel nonmyopic ϵ-Bayes-optimal active learning (ϵ-BAL) approach that jointly optimizes the trade-off. In contrast, existing works have primarily developed greedy algorithms or performed exploration and exploitation separately. To perform active learning in real time, we then propose an anytime algorithm based on ϵ-BAL with performance guarantee and empirically demonstrate using a real-world dataset that, with limited budget, it outperforms the state-of-the-art algorithms.