Active Learning Policies for Solving Inverse Problems

We propose to learn an active learning policy for inverse problem optimisation.

Switching policies for solving inverse problems

We propose to learn a switching policy for inverse problem optimisation.

Learning objective-specific active learning strategies with Attentive Neural Processes

We propose a novel learning-active-learning method that can adapt to different task objectives.

E-Valuating Classifier Two-Sample Tests

We propose an E-values based classifier two-sample test that has much stronger finite-sample type-I error control than existing methods.