Exploring the Design Dimensions of Curiosity for Robots

Due to the advances in Artificial Intelligence technologies and Robotics, the field of Human-Robot Interaction is currently able to envision robots as constant collaboration partners for a magnitude of daily tasks. With robots being exposed to different environments and human needs, they need to be able to learn multiple different personalized tasks. Currently, this is mostly done with Robot Learning from Demonstration. However, due to the sheer amount of possible tasks, objects, and interactions, it is an impossible task to teach robots every subtask manually. Thus, we propose that enabling curiosity for robots; giving them an intrinsic desire to learn.

In this thesis, we will investigate the human aspect of this learning process. Teaching other humans new skills is a hard job, but humans clearly see a long-term benefit. Thus, humans are kind and supportive and therefore, also show a new skill multiple times. Will this be the same for training our domestic robot? For example, showing them how to cook? or simply how to open a drawer? Moreover, when giving the robot humanoid characters are humans more likely to put more effort into training their robot?

You will:

• Perform a literature review
• Identify possible design dimensions
• Develop a study apparatus using the dimensions
• Conduct an evaluation study
• Summarize your findings in a thesis and presenting them to an audience
• (Optional) Co-writing a research paper

You need:

• Strong communication skills in english
• Good knowledge of study design
• Good knowledge of data analysis
• Good knowledge of statistical evaluation
• (Optional) ROS skills

References

• Automatic Class Discovery and One-Shot Interactions for Acoustic Activity Recognition