Cognitive Modeling in Robot Learning for Adaptive Human-Robot Interactions


Our interdisciplinary workshop focuses on the human-centered design of adaptive robotic behavior from the lens of cognitive science. In this workshop, we invite researchers from different backgrounds, including engineering, human-machine interaction, and cognitive science, to discuss how cognitive modeling can be exploited to enhance adaptive human-robot interaction (HRI) frameworks. We aim to provide participants diverse theoretical perspectives and potential research directions through interactive talks and panel discussions. 



Over the past two decades, there has been increased research interest in the Human-Robot Interaction (HRI) field: scientific papers published in HRI raised from about a hundred papers to almost ten thousand in 2020. However, this increased research interest does not comprehensively integrate the concepts of human-centered design (HCD) and thus does not utilize the full capabilities of HRI-based frameworks since less than 30% of the papers published in 2020 developed HRI frameworks focused on HCD concepts. HCD is important for better HRI as real-life applications often require humans and robots to cognitively align goals, have a shared task representation, and adapt to each other’s behaviors. The low amount of work focusing on HCD design emphasizes the need to promote its inclusion in HRI research-an interdisciplinary research domain by nature. To address this issue, we propose an interdisciplinary workshop focusing on the aspects of HCD from the lens of cognitive science. In this workshop, we discuss how cognitive modeling can be applied for adaptive human-robot interaction from diverse theoretical perspectives and research domains by bringing speakers from engineering, human-machine interaction research, and cognitive science.



HRI can be implemented in different ways—either the human and the robot can collaborate in close contact (e.g., coordinated lifting tasks), human-robot cooperation tasks (e.g., human and the robot work alternately on different tasks within a process without direct interaction but share the same objective and workspace), or the human and the robot can interact remotely (e.g., teleoperation of a robotic system with the help of computer applications such as virtual reality), where the robot assists the human in tasks deemed too dangerous for direct human involvement or in tasks in hard-to-reach places or hostile environments. Both types of HRI require the human agent and the robotic system to adapt to each and the interaction environment.

Recent cognitive science and computational modeling works can inform adaptive HRI for robotics. For example, in an HRI task, given observed human behaviors, with considerations for the human’s cognitive bounds and the task’s environmental bounds, these computational models can help the robot infer the human’s goals, intentions, or even the subjective utility functions. In addition, these models can also help the robot predict human decisions and behaviors given the inferred goals. Such considerations of the human agent will be more likely to reduce the inconvenience, threat, annoyance, or harm to human users and provide further accessibility, functionality, and protection instead. This approach to improving the quality of human-robot interaction aligns with the goal of bringing a human-centered design approach to the robotic systems’ development processes. 

Given this background and motivation, our workshop focuses on three main questions: 

  1. How can human-centered design improve human-robot interactions?
  2. How can cognitive models help develop robotic strategies?
  3. Why is adaptive human-robot learning important, and how can we model adaptive human-robot interaction?



Topics of interest include but are not limited to

  • Architectures and frameworks for cognition
  • Cognitive human–robot interaction
  • Cognitive load in HRC
  • Cognitive modeling and development
  • Cognitive robotics
  • Human and robotic action-effect learning
  • Knowledge discovery and representation in robots
  • Knowledge representation and reasoning
  • Learning for action and interaction
  • Neurorobotics
  • New paradigms and perspectives
  • Psychological aspect of cognitive robotics
  • Robots to identify problems and questions about cognition
  • Safety in collaborative robotics
  • Social and assistive robots


Invited Speakers:

  1. Dr. Anca Dragan,
    Associate Professor, Electrical Engineering and Computer Sciences, UC Berkeley, USA.
    Topic: Cognitive Models Beyond Noisy-Rationality
  2. Dr. Henny Admoni,
    Assistant Professor, Human And Robot Partners Lab, Robotics Institute, Carnegie Mellon University, USA
    Topic: Modeling Intent Through Nonverbal Behaviors in Human-Robot Interaction
  3. Dr. Andrew Howes,
    Professor & Head of School, School of Computer Science, University of Birmingham, UK
    Topic: Models of Human Cognition for Artificial Intelligence
  4. Dr. Yukie Nagai,
    Project Professor, University of Tokyo
    Topic: Predictive Coding as a Unified Theory for Human and Robot Cognition
  5. Dr. Arash Ajoudani,
    Senior Scientist, Italian Institute of Technology, Italy
    Topic: Personalizable Human-Robot Interaction Through Cognitive Load Monitoring and Optimization
  6. Dr. Christoforos Mavrogiannis,
    Assistant professor, University of Michigan, USA
    Topic: Building robots that humans accept


Tentative Schedule:

We plan to organize a full-day workshop, and below is a tentative schedule. Each 45 min talk is divided into 30 minutes for the talk itself, 10 minutes for Q&A/discussion, and 5 minutes for speaker change and setup.

Time Description
08h30 – 09h00 Welcome
09h00 – 09h45 Invited talk 1 – Dr. Anca Dragan
09h45 – 10h30 Invited talk 2 – Dr. Christoforos Mavrogiannis
10h30 – 10h55 Coffee break
10h55 – 11h40 Invited talk 3 – Dr. Henny Admoni
11h40 – 12h25 Invited talk 4 – Dr. Andrew Howes
12h25 – 13h30 Lunch break
13h30 – 14h15 Invited talk 5 – Dr. Yukie Nagai
14h15 – 15h00 Invited talk 6 – Dr. Arash Ajoudani
15h00 – 15h20 Spotlight talks from submitted abstracts
15h00 – 15h45 Coffee break + Posters
15h45 – 16h30 Panel discussion with speakers
16h30 – 16h40 Final remarks



Anany Dwivedi, Chenxu Hao, Gustavo J. G. Lahr, Marta Lorenzini

The workshop is proposed for IEEE International Conference on Robotics and Automation (ICRA), 2023 at London, UK.