Towards Physiologically Sensible Predictions via the Rule-based Reinforcement Learning Layer

TL;DR

This paper introduces a lightweight, rule-based reinforcement learning layer that corrects physiologically impossible predictions in healthcare models, significantly improving their accuracy without requiring extensive expert input.

Contribution

The novel RRLL framework effectively reduces impossible healthcare predictions by using a small set of rules, enhancing existing models without heavy domain knowledge.

Findings

RRLL reduces physiologically impossible predictions

Significant accuracy improvements across multiple healthcare tasks

Efficient correction with minimal expert input

Abstract

This paper adds to the growing literature of reinforcement learning (RL) for healthcare by proposing a novel paradigm: augmenting any predictor with Rule-based RL Layer (RRLL) that corrects the model's physiologically impossible predictions. Specifically, RRLL takes as input states predicted labels and outputs corrected labels as actions. The reward of the state-action pair is evaluated by a set of general rules. RRLL is efficient, general and lightweight: it does not require heavy expert knowledge like prior work but only a set of impossible transitions. This set is much smaller than all possible transitions; yet it can effectively reduce physiologically impossible mistakes made by the state-of-the-art predictor models. We verify the utility of RRLL on a variety of important healthcare classification problems and observe significant improvements using the same setup, with only the domain-specific set of impossibility changed. In-depth analysis shows that RRLL indeed improves accuracy by effectively reducing the presence of physiologically impossible predictions.