Lingodroids
Lingodroids are language learning robots that play location language games to construct shared lexicons for places, distances, and directions.
Lingodroids playing various location language games Video
Toponymic Lexicon:
The robots play where-are-we games, in which they name the current location, to form a toponymic lexicon, where toponyms are names for places.
e.g. “jaya” = “At the centre of the room”
Testing a Toponymic Lexicon:
They can test their toponymic lexicon by playing go-to games, in which they attempt to meet at a distant location.
Generative Games:
Once they have a toponymic lexicon, they can expand their concepts by playing generative games to form concepts for distances and directions.
e.g. “kuzo reya duka” = “The bottom left corner is nearby to the bottom right corner”
e.g. “kuzo pize reya hiza” = “If I’m at the bottom left corner facing the top left corner then the bottom right corner is to the right of me”
With distances and directions, they can now expand the places they can talk about, and actually meet at these new locations, should they become accessible, for example, if office doors are opened.
e.g. “reya rije duka hiza heto” = “If I’m at the bottom right corner of the room facing the top right corner of the room, then nearby to the right is a place that I can’t get to that I’m calling heto”
Images:
For more information:
Schulz, R., Glover, A., Milford, M., Wyeth, G., & Wiles, J. (2011) Lingodroids: Studies in Spatial Cognition and Language, ICRA 2011, The International Conference on Robotics and Automation, Shanghai, China, May 2011
The Lingodroids are a pair of mobile robots that evolve a language for places and relationships between places (based on distance and direction). Each robot in these studies has its own understanding of the layout of the world, based on its unique experiences and exploration of the environment. Despite having different internal representations of the world, the robots are able to develop a common lexicon for places, and then use simple sentences to explain and understand relationships between places – even places that they could not physically experience, such as areas behind closed doors. By learning the language, the robots are able to develop representations for places that are inaccessible to them, and later, when the doors are opened, use those representations to perform goal-directed behavior.
Funded by:
2009-2012 Talking with Robots: Evolving Grounded Language for Embodied Agents, ARC Discovery Grant to Professor Janet Wiles and Professor Gordon Wyeth.