Designing an Interpretable Interface for Contextual Bandits

TL;DR

This paper introduces an interpretable interface for contextual bandits that helps non-expert operators understand and manage these systems effectively, using a novel metric called 'value gain' and validated through user studies.

Contribution

The paper presents a new interface design for contextual bandits that enhances interpretability for non-experts, incorporating the 'value gain' metric and qualitative evaluation.

Findings

The interface improves non-experts' understanding of bandit behaviour.

Balancing technical detail with accessibility empowers users to manage systems.

The 'value gain' metric effectively quantifies sub-component impacts.

Abstract

Contextual bandits have become an increasingly popular solution for personalized recommender systems. Despite their growing use, the interpretability of these systems remains a significant challenge, particularly for the often non-expert operators tasked with ensuring their optimal performance. In this paper, we address this challenge by designing a new interface to explain to domain experts the underlying behaviour of a bandit. Central is a metric we term "value gain", a measure derived from off-policy evaluation to quantify the real-world impact of sub-components within a bandit. We conduct a qualitative user study to evaluate the effectiveness of our interface. Our findings suggest that by carefully balancing technical rigour with accessible presentation, it is possible to empower non-experts to manage complex machine learning systems. We conclude by outlining guiding principles that other researchers should consider when building similar such interfaces in future.