Classification with Costly Features as a Sequential Decision-Making Problem
This addresses a practical problem for applications like medical diagnosis or finance where feature acquisition is costly, though it is an incremental improvement over existing sequential decision-making approaches.
The paper tackles the classification problem where features must be acquired at a cost under a per-sample budget, converting it into a Markov decision process solved with deep reinforcement learning. The method outperforms prior-art algorithms across multiple datasets and settings, including sparse training data with missing features.
This work focuses on a specific classification problem, where the information about a sample is not readily available, but has to be acquired for a cost, and there is a per-sample budget. Inspired by real-world use-cases, we analyze average and hard variations of a directly specified budget. We postulate the problem in its explicit formulation and then convert it into an equivalent MDP, that can be solved with deep reinforcement learning. Also, we evaluate a real-world inspired setting with sparse training dataset with missing features. The presented method performs robustly well in all settings across several distinct datasets, outperforming other prior-art algorithms. The method is flexible, as showcased with all mentioned modifications and can be improved with any domain independent advancement in RL.