Fork the consciousness, or download the project and create your own.

Jürgen Schmidhuber

IDSIA Lugano; KAUST AI Initiative

Integrated information Consciousness is intrinsic causal power, measured as integrated information

Jürgen Schmidhuber, co-director of the Dalle Molle Institute for Artificial Intelligence Research (IDSIA) in Lugano and head of the AI Initiative at KAUST, is among the most consequential living machine learning researchers, with foundational contributions to recurrent networks, meta-learning, and curiosity-driven exploration. His engagement with consciousness is architectural rather than metaphysical. His self-referential weight matrices and Gödel machine designs give an agent a machine-learning model of its own computational process that the agent uses to improve itself, and his essay series on learning to think argues that the interesting question is how a system comes to contain and exploit a model of itself.

The position that follows is a strict computational one. Whatever consciousness is, on his account the scientifically tractable part is self-modeling and self-improvement in general problem solvers, and the computable universe he describes admits artificial systems whose self-models are as rich as ours. He states plainly that he considers the consciousness question badly posed, and prefers to build the machines whose self-referential structure makes the question concrete.

For this site, Schmidhuber matters as the boundary case. His self-modeling machines supply exactly the architectural substrate theories of machine consciousness describe, while declining every phenomenal claim, which keeps the distinction between building the structure and explaining the experience in view.

Known for. Self-referential weight matrices, the Gödel machine, learning to think, computable universes, meta-learning machines that model their own search

Coverage on this site

No articles yet. This profile exists because the researcher's work is referenced elsewhere on the site, and coverage will be linked here as it is published.

Related researchers