The Autodidactic Universe

• DOI: 10.48550/arxiv.2104.03902
• Publication Date: 2021-01-01
• Authors: Stephon Alexander; W. J. Cunningham; Jaron Lanier; Lee Smolin; Stefan Stanojevic; Michael W. Toomey; Dave Wecker
• Keywords: Physical system; Matrix (chemical analysis)

We present an approach to cosmology in which the Universe learns its own physical laws. It does so by exploring a landscape of possible laws, we express as certain class matrix models. discover maps that put each these models correspondence with both gauge/gravity theory and mathematical model learning machine, such deep recurrent, cyclic neural network. This establishes between solution run is not equivalence, partly because gauge theories emerge from $N \rightarrow \infty $ limits models, whereas same networks used here are well-defined. discuss detail what it means say takes place autodidactic systems, where there no supervision. propose if network can be said learn without supervision, for corresponding theory. consider other protocols optimization graph variety, subset-replication using self-attention look-ahead, geometrogenesis guided reinforcement learning, structural renormalization group techniques, extensions. These together provide number directions explore origin laws based on putting machine architectures theories.