# Learning to Acquire Information

## TL;DR

This paper introduces an active diagnosis method that predicts the most informative next observation using an implication model, reducing computational complexity and learning query strategies without needing to understand the hidden hypothesis.

## Contribution

It proposes a novel active diagnosis approach based on an implication model that predicts observation outcomes and selects high-entropy observations, bypassing complex hypothesis inference.

## Key findings

- Reduces computational complexity in diagnosis tasks.
- Learns to select observations without explicit hypothesis modeling.
- Effective in scenarios with uniform observation entropy.

## Abstract

We consider the problem of diagnosis where a set of simple observations are used to infer a potentially complex hidden hypothesis. Finding the optimal subset of observations is intractable in general, thus we focus on the problem of active diagnosis, where the agent selects the next most-informative observation based on the results of previous observations. We show that under the assumption of uniform observation entropy, one can build an implication model which directly predicts the outcome of the potential next observation conditioned on the results of past observations, and selects the observation with the maximum entropy. This approach enjoys reduced computation complexity by bypassing the complicated hypothesis space, and can be trained on observation data alone, learning how to query without knowledge of the hidden hypothesis.