AI Researcher To Build…AI Researcher

Artificial Intelligence researcher Jurgen Schmidhuber says his main scientific ambition “is to build an optimal scientist, then retire!”

The Cognitive Robotics professor has worked on problems including artificial ants and even robots that are taught how to tie shoelaces using reinforcement learning, but he believes algorithms can be written that allow the programming of curiosity itself. And he offers a fascinating metaphor for life after the development of AI.”It’s a bit like asking an ant of 10 million years ago: If humans were created tomorrow, what sort of implications do you think that would have for all the ant colonies?

Jürgen Schmidhuber at Singularity Summit 2009 – Compression Progress: The Algorithmic Principle Behind Curiosity and Creativity from Michael Anissimov on Vimeo.

“In hindsight we know that many ant colonies are still doing fine, but some of them (for example, those in my house) have goal conflicts with humans, and live dangerously.”

He’s also created art using algorithmic information theory, and describes the simple algorithmic principle that underlies subjective beauty and creativity. “All we need is (1) An adaptive predictor or compressor of the continually growing sensory data history… (2) A learning algorithm that continually improves the predictor or compressor… (3) Intrinsic rewards measuring the predictor’s or compressor’s improvements due to the learning algorithm, and (4) A reward optimizer or reinforcement learner that translates those rewards into action sequences expected to optimize future reward!”

And ultimately, he addresses the possibility that the entire Universe, including everyone in it, is in principle computable by a completely deterministic computer program!

, ,

  • subpixel

    He may be neglecting notions such as the uncertainty principle; the current state of the universe cannot be determined with total precision, thus the next state of the universe cannot be precisely derived from it.

  • subpixel

    He may be neglecting notions such as the uncertainty principle; the current state of the universe cannot be determined with total precision, thus the next state of the universe cannot be precisely derived from it.

  • Marc

    What could go wrong?

  • Marc

    What could go wrong?

21