Learning to Maximize Mutual Information for Dynamic Feature Selection
This work addresses the problem of reducing data acquisition costs in machine learning for practitioners by offering a simpler alternative to reinforcement learning for dynamic feature selection, though it is incremental in nature.
The paper tackles the dynamic feature selection problem by proposing a method that greedily selects features based on conditional mutual information, using amortized optimization to learn the policy without oracle access. It outperforms existing methods in experiments, validating it as a simple yet effective approach.
Feature selection helps reduce data acquisition costs in ML, but the standard approach is to train models with static feature subsets. Here, we consider the dynamic feature selection (DFS) problem where a model sequentially queries features based on the presently available information. DFS is often addressed with reinforcement learning, but we explore a simpler approach of greedily selecting features based on their conditional mutual information. This method is theoretically appealing but requires oracle access to the data distribution, so we develop a learning approach based on amortized optimization. The proposed method is shown to recover the greedy policy when trained to optimality, and it outperforms numerous existing feature selection methods in our experiments, thus validating it as a simple but powerful approach for this problem.