An Information-Theoretic Framework for Unifying Active Learning Problems
Keywords:Active Learning, Algorithm Configuration, Human-in-the-loop Machine Learning
AbstractThis paper presents an information-theoretic framework for unifying active learning problems: level set estimation (LSE), Bayesian optimization (BO), and their generalized variant. We first introduce a novel active learning criterion that subsumes an existing LSE algorithm and achieves state-of-the-art performance in LSE problems with a continuous input domain. Then, by exploiting the relationship between LSE and BO, we design a competitive information-theoretic acquisition function for BO that has interesting connections to upper confidence bound and max-value entropy search (MES). The latter connection reveals a drawback of MES which has important implications on not only MES but also on other MES-based acquisition functions. Finally, our unifying information-theoretic framework can be applied to solve a generalized problem of LSE and BO involving multiple level sets in a data-efficient manner. We empirically evaluate the performance of our proposed algorithms using synthetic benchmark functions, a real-world dataset, and in hyperparameter tuning of machine learning models.
How to Cite
Nguyen, Q. P., Low, B. K. H., & Jaillet, P. (2021). An Information-Theoretic Framework for Unifying Active Learning Problems. Proceedings of the AAAI Conference on Artificial Intelligence, 35(10), 9126-9134. Retrieved from https://ojs.aaai.org/index.php/AAAI/article/view/17102
AAAI Technical Track on Machine Learning III