Learning Algorithms for Active Learning
This work addresses the challenge of automating active learning algorithm design for machine learning practitioners, though it appears incremental as it builds on existing metalearning and active learning concepts.
The authors tackled the problem of designing active learning algorithms by introducing a metalearning model that jointly learns data representation, item selection heuristics, and prediction function construction for related tasks, demonstrating its effectiveness on Omniglot and MovieLens datasets in synthetic and practical settings.
We introduce a model that learns active learning algorithms via metalearning. For a distribution of related tasks, our model jointly learns: a data representation, an item selection heuristic, and a method for constructing prediction functions from labeled training sets. Our model uses the item selection heuristic to gather labeled training sets from which to construct prediction functions. Using the Omniglot and MovieLens datasets, we test our model in synthetic and practical settings.