The Babybot Challenge
The field of developmental robotics forms a bridge between two research communities: those who study learning and development in humans and those who study comparable processes in artificial systems. The Babybot Challenge was designed to help strengthen this bridge. Specifically, participants in the challenge will select from a list of three infant studies, and design a model that captures infants’ performance on the chosen task. Submissions will be judged by three criteria: (1) How well does the model represent the particular features of the research paradigm? (2) How closely does the performance of the model replicate the findings from the chosen study? And (3) what novel insights or explanations for the observed developmental pattern are generated by the model?
June 15, 2015 (note that this is later than the deadline for standard paper submissions)
1st prize = $500, 2nd prize = $300, 3rd prize = $200
Participation in the challenge is divided into two stages. During the first stage of the challenge, authors will submit a 4-8-page paper through ras.paperplaza.net (there is no additional charge for pages 7/8). Papers will be pre-screened, and submissions that are accepted for the challenge will then participate in the second stage, an oral presentation at a special session of the conference. Additional instructions for submitting the paper, including details on the judging process, will be posted soon.
Infant Studies (please select one)
Cognition/Perception: Sommerville, J.A., Woodward, A.L., & Needham, A. (2005) Action experience alters 3-month-old infants’ perception of others’ actions. Cognition, 96, B1-B11. (pdf)
Language/Social: Iverson, J. M., & Goldin-Meadow, S. (2005). Gesture paves the way for language development. Psychological Science, 16, 367-371. (pdf)
Motor Skill: von Hofsten, C. (1984). Developmental changes in the organization of prereaching movements. Developmental Psychology, 20, 378-388. (pdf)
Can the model focus exclusively on the neural/information processing in the task, without an embodied aspect?
The primary goal of the Babybot Challenge is to use modeling as a tool to generate meaningful insights about how development occurs on the chosen task or paradigm. Whether a model does this through representing the child as an embodied agent, or through a more abstract representation, is less critical. Note that while submissions that address both the neural and information processing aspects of the task as well as the embodied features of the task are expected to better meet the judging criteria, submissions of any type are welcome. Please email us (see Contact above) if you have questions about this!
Does the model need to be implemented on a real robot platform?
Absolutely not - feel free to do it in simulation - write your own code, use an existing simulation software package, or any combination of the two. However, during the second stage of the competition (oral presentation), submissions should ideally include either a recorded or “real-time” demonstration of how the simulated child performs in the study.
How much of the infant study that I selected needs to be replicated by the model?
Submissions will be judged on how well they capture some core findings from the infant study - it’s not necessary to reproduce the paper in full detail.