Any-Cubes ¨C Discover Artificial Intelligence through Playing

The aim of the research and development project "Any-Cubes" is to make it as easy as possible for children to play and experiment with artificial intelligence and networked objects.

To this end, a functioning prototype was developed with which children can discover artificial intelligence (machine learning) and networked objects step by step and integrate them into their play. The central pedagogical starting point of the project is that children themselves are the actors of their play. Children use toys as play material. A good learning toy therefore enables creativity, imagination and cooperation and can be freely linked with other forms of play and materials. An important research question of Any-Cubes is therefore: How must a computational toy be designed so that it can serve children as a universal tool and stimulate experimentation and creation?

Technically, Any-Cubes is a combination of two future-oriented topics. On the one hand, deep learning-based image classification and, on the other, the networking of physical objects (Internet of Things). The set consists of a sensor cube and a maker cube. The sensor cube can be trained for every possible camera image at the touch of a button. If the Sensor-Cube detects a previously learned object, it sends a signal to the Maker-Cube. The Maker-Cube offers up to four switches to which small toys (with standard household battery voltage) can be connected. This allows children to create simple if-then functions and integrate them into their play, regardless of location and in real time. The machine learning functionality is realised directly on a developer board using existing open source libraries. The data is thus only processed locally and not sent over the internet.

In 2019, Any-Cubes was awarded an Honorable Mention Demo Award at the Mensch und Computer conference in Hamburg.

Any-Cubes is a research project of the Department of Education and Social Sciences at FH Potsdam and is funded by the Research and Development Fund of FH Potsdam.

For further information, please send an email to scheidt@fh-potsdam.de.

Publication
Scheidt, A., & Pulver, T. (2019). Any-Cubes: A Children's Toy for Learning AI. Proceedings of Humans and Computers 2019, 893-895. doi: 10.1145/3340764.3345375