Home Speakers Schedule Call for papers Accepted papers Organisers Highlights
08:45 Welcome and opening remarks
09:00 Dongheui Lee (TU Wien)
09:30 Shuran Song (Stanford University)
10:00 Coffee break + poster session
10:30 Chris Pek (TU Delft)
Please Don't Fail Me Again: Tackling Robustness in Task and Motion Planning
11:00 Spotlight presentations of accepted papers
11:30 Challenges in handling execution failures in human-robot interaction and collaboration
Panel discussion with the invited speakers
12:30 Lunch break
14:00 Alessandra Sciutti (Italian Institute of Technology)
Failures and Trust in Human-Robot Interaction
Effective human-robot interaction (HRI) hinges on understanding and managing trust, especially amidst robotic unreliability. This talk examines how a robot's hardware and social failures influence interaction dynamics and trust toward a robotic partner. Moreover, it will delve into the human biases that might lead to overtrusting a robot despite its failures or not relying on its expertise - in the face of its high performance. This presentation emphasizes the importance of understanding human a priori expectations about robot capabilities and their responses to robotic failures to manage trust effectively and develop adaptive and reliable robotic systems.
14:30 Maartje de Graaf (Utrecht University)
Understanding and Mitigating Interactive Breakdowns in Human-Robot Collaboration
Because of the cognitive differences between people and robots, interactive breakdowns are unavoidable in human-robot interactions. People's incomplete knowledge of robot capabilities can lead them to believe that robot behaviors are erroneous, even when they are following preprogrammed tasks or actions. Using a multidisciplinary approach based on social psychology, communication science, and social robotics, my research centers on elucidating human perceptions of erroneous robots to develop interactive strategies to equip robots with coping mechanisms for such scenarios, with applications to collaborative tasks, hospitality, and child healthcare. In this talk, I will discuss research findings on the psychological mechanisms influencing human perception of robot behavior, as well as strategies for robot interventions to restore trust following interactive breakdowns.
15:00 Maximilian Diehl (Chalmers University of Technology)
Explaining and Preventing Robot Failures through Causal Models
Future robots are expected to assist humans with daily tasks such as setting the table and washing dishes. However, operating in human environments, these robots are prone to errors. Humans often find it challenging to understand why these failures occur, especially when robots rely on black-box decision-making methods. This lack of transparency reduces trust and effectiveness in human-robot interactions and limits humans' ability to assist robots in recovering from failures. In this talk, I will present our methods that use causal models of robot actions to explain why these failures occur. I will discuss different ways of learning cause-effect relationships between task executions and their outcomes, how we use these causal models to provide contrastive failure explanations, and how to predict and prevent future failures.
15:30 Coffee break + poster session
16:00 Challenges in handling execution failures in human-robot interaction and collaboration
Interactive session: During the interactive session, participants will gather into groups to discuss concrete aspects of execution failures in HRI and HRC, such as (i) types of failures that are commonly encountered in deployments of robots in HRI and HRC scenarios, (ii) the effect of failures on robot trustworthiness, or (iii) considerations for collecting robot training data that represent execution failures.
16:45 Closing remarks