Artificial Intelligence via Telepresence

Most research and application in the field of robotics has attempted to instill automation intelligence into robots via designer algorithms. Attempting to create intelligence behavior in this manner is akin to running up a cliff. While some progress can be made, the heights remain far out of reach. Modern approaches mix in neural networks, genetic algorithms, and a variety of other learning systems. Learning systems are a major leap forward in bringing intelligent behavior to machines, but these methods are generally hampered by limited input, feedback and achievable goals.

I propose we take a more synergetic approach between man and machine, to allow the robots to read our “thoughts”, so to speak, and in accordance, do our bidding. Initially a robot’s “host” would control the movement of a robot’s limbs through a mind and body interfaces. Microsoft’s recent Natal interface is a prefect example. Then, using pattern recognition networks, such as Bayesian networks and neural networks, the robots would be able to generalize on this human input, learning to preform some specific task. These action patterns could then be encapsulated as unified behaviors, such as command to a robotic dog, “fetch paper”. Combining these individual trainings into larger systems could then produce very intelligent behaviors, vastly superior to any designs we have been able to implement manually.

Having the robot under self-control, readily available for general use to perform specialized tasks, is the ultimate achievement of industrial automation. It behooves us to put our own physical skills to work as a means of education for automated learning systems, rather than attempting to instill the behaviors or criteria via our crude means of symbolic programming.

Tags: ai, robots