Is Information Digital? A Defense of Reality

Data-driven techniques such as large language models have shown remarkable progress in recent years. In contrast, advances in cyber-physical systems — including robotics and autonomous machines — have been significantly slower. In his lecture, Prof. Edward A. Lee explores whether this difference may be rooted in a more fundamental issue: the limits of digital representation and objective observation.

Drawing on concepts from computer science such as zero-knowledge proofs and bisimulation, and connecting them to Shannon information theory, Lee argues that objective observation may reveal only a limited subset of the information embedded in physical reality. Some aspects of the real world, he suggests, cannot be fully represented digitally. If this is the case, learning — even for machines — may require embodied interaction with the physical world, much like learning to ride a bicycle requires actually getting on one.

Edward A. Lee is Distinguished Professor Emeritus in Electrical Engineering and Computer Sciences at the University of California, Berkeley. His research focuses on cyber-physical systems and the integration of physical processes with software and networks.

More information

 Contact person for queries: