My research aims to contribute to developing intelligent agent algorithms that are useful and enjoyable to collaborate with when making decisions. For my dissertation, I look at the possible tension between performance goals (e.g. decision quality, efficiency, individual goals) and social goals (e.g. respecting human agency, maintaining trust, team goals) in collaborative decision-making. I prototype agents with differing priorities to highlight the trade-offs and offer design insights on how future agents can effectively balance the two goals. [dissertation]
A VR system in which three participants interact at an ambiguous all-way stop intersection simultaneously, assuming the roles of a pedestrian, a driver, and a passenger in an autonomous vehicle(AV) . Should AVs be more proactive to get to a destination or focus on yielding to others? How should they manage different expectations from different road users?
Unity , C# , Body-tracking , Virtual Reality
A two-player search-and-rescue game as a sequential social dilemma that highlights the tension between individual and team goals. How do human players use their social intelligence and skills to balance the two goals? How might we develop an artificial social intelligence that balance its individual goals with the team's goals?
Humans sharing agency with a virtual robot to jointly arrive at a decision for a preference-based ranking task. Should robots purely focus on decision quality or suggest a compromise with human decisions when they disagree?