Bridging the Gap in Hybrid Decision-Making Systems
Federico Mazzoni, Roberto Pellungrini, Riccardo Guidotti
TL;DR
This work addresses the challenge of balancing human and machine control in high-stakes hybrid decision-making by introducing BRIDGET, a two-phase, co-evolutionary system that blends Learning-to-Defer and Skeptical Learning within an Incremental Learning framework. BRIDGET uses a Human-in-Command and a Machine-in-Command mode, with fading skepticality and fading empirical accuracy to measure reliability and guide phase transitions, while providing explanations to maintain interpretability. The key contributions are the formalization of a co-evolutionary human–machine framework that can switch leadership dynamically and the integration of explainable guidance to support user trust and continual learning. The approach holds potential for flexible, interpretable collaboration in domains requiring ongoing human oversight and machine assistance, with mechanisms to detect drift and novelty and adapt accordingly.
Abstract
We introduce BRIDGET, a novel human-in-the-loop system for hybrid decision-making, aiding the user to label records from an un-labeled dataset, attempting to ``bridge the gap'' between the two most popular Hybrid Decision-Making paradigms: those featuring the human in a leading position, and the other with a machine making most of the decisions. BRIDGET understands when either a machine or a human user should be in charge, dynamically switching between two statuses. In the different statuses, BRIDGET still fosters the human-AI interaction, either having a machine learning model assuming skeptical stances towards the user and offering them suggestions, or towards itself and calling the user back. We believe our proposal lays the groundwork for future synergistic systems involving a human and a machine decision-makers.