Jiali Duan

About

Jiali Duan’s research sits at the intersection of robotics, human-robot interaction, and adversarial machine learning, with a focus on rethinking how robots learn from people. Her most cited work, “Robot Learning via Human Adversarial Games” (2019, 4 citations), challenges the prevailing assumption that human supervisors are always cooperative. Instead, Duan models human observers as potentially adversarial agents—introducing a game-theoretic framework where robots must learn robust policies despite uncooperative or even deceptive human behavior. This contribution is significant because it bridges a critical gap between idealized human-in-the-loop systems and the messy realities of deployment, where human intentions may be ambiguous or conflicting. By reframing the learning process as an adversarial game, Duan’s work has implications for safer, more resilient autonomous systems in applications like collaborative manufacturing, assistive robotics, and autonomous driving. Though her citation count is modest, the conceptual novelty of her approach has sparked interest in adversarial robustness for interactive learning. Duan’s research is a compelling call to design robots that can handle not just cooperative partners, but the full spectrum of human behavior.

Research Focus

Key Achievements

Top Papers

1. 1
   Robot Learning via Human Adversarial Games

   4 citations · 2019
2. 2
   Robot Learning via Human Adversarial Games

   3 citations · 2019

Key Collaborators

Contact & Links