Diagnosis, Feedback, Adaptation: A Human-in-the-Loop Framework for Test-Time Policy Adaptation

TL;DR

This paper introduces a human-in-the-loop framework that uses user feedback and counterfactual demonstrations to identify task-irrelevant concepts, enabling personalized policy adaptation and improved robustness in control tasks.

Contribution

It presents a novel interactive approach combining feedback and counterfactuals for personalized task-irrelevant concept identification and policy adaptation.

Findings

Reduces demonstrations needed for fine-tuning

Improves understanding of agent failure modes

Aligns policies with individual user preferences

Abstract

Policies often fail due to distribution shift -- changes in the state and reward that occur when a policy is deployed in new environments. Data augmentation can increase robustness by making the model invariant to task-irrelevant changes in the agent's observation. However, designers don't know which concepts are irrelevant a priori, especially when different end users have different preferences about how the task is performed. We propose an interactive framework to leverage feedback directly from the user to identify personalized task-irrelevant concepts. Our key idea is to generate counterfactual demonstrations that allow users to quickly identify possible task-relevant and irrelevant concepts. The knowledge of task-irrelevant concepts is then used to perform data augmentation and thus obtain a policy adapted to personalized user objectives. We present experiments validating our framework on discrete and continuous control tasks with real human users. Our method (1) enables users to better understand agent failure, (2) reduces the number of demonstrations required for fine-tuning, and (3) aligns the agent to individual user task preferences.