How about a robotic baby?
Artificial intelligence is an attempt to imitate the reasoning human behavior by machine. Unlike computer, designed to process steps by steps, human reasoning is less structured. Artificial intelligence is therefore designed to follow these processes.
Artificial neural networks are example of machine learning with reasoning capabilities. They are computational models with the abilities such as adapt, learn, generalize, recognize, cluster, and organize data. These are neural computing using iterative processing system to function like a human brain.
A report from New Scientists revealed that Babybot created by roboticists from Italy, France and Switzerland, experiments with objects nearby and learns from them. This provides the robot an ability to develop motor skills like a human infant.
The “brain” of the robot baby is actually a cluster of 20 computers running several neural networks. The software copies a biological neural system and learns by establishing and altering the strength of links between artificial neurons. The neural network software can be adjusted to observe the robot’s learning behavior allowing the robotics to test different neuroscience models. The aim is to build a humanoid 2-year old child with all of Babybot’s abilities and it may eventually design to learn how to crawl and walk.
However, researcher believes fundamental differences between the human brain and computers used to control learning robots like Babybot may mean that such machines can never become as intelligent as us.