• Corpus ID: 233181473

The Autodidactic Universe

  title={The Autodidactic Universe},
  author={Stephon H. S. Alexander and William J. Cunningham and Jaron Lanier and Lee Smolin and Stefan Stanojevic and Michael W Toomey and Dave Wecker},
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, which we express as a certain class of matrix models. We discover maps that put each of these matrix models in correspondence with both a gauge/gravity theory and a mathematical model of a learning machine, such as a deep recurrent, cyclic neural network. This establishes a correspondence between each solution of the physical theory and a run of a neural… 
A Use of Variety as a Law of the Universe
This paper explores the idea of a toy model universe as a character string where each character is either X or -. The string makes a transition between states that are of maximal variety. The
A Cubic Matrix Action for the Standard Model and Beyond
A new framework for matrix theories that are equivalent to field theories on a toroidal spacetime are proposed, accomplished via infinite Toeplitz matrices whose entries match the field degrees of freedom on an energy-momentum lattice.
Towards a Theory of Quantum Gravity from Neural Networks
It is argued that the Lorentz symmetries and curved space-time can emerge from the interplay between stochastic entropy production and entropy destruction due to learning.
The Communication Infrastructures of Living Systems and the Universe: Analysis from the Perspective of Antenna Theory
It is still unknown how molecules coordinate their activity and operate at high speeds in the crowded environment of a cell. The study focuses on the geometry of biomolecules, assuming B-DNA,
The Physics of Machine Learning: An Intuitive Introduction for the Physical Scientist
A review of two energy-based machine learning algorithms, Hopfield networks and Boltzmann machines, and their connection to the Ising model serves as a foundation to understand the phenomenon of learning more generally.
Geometric State Sum Models from Quasicrystals
In light of the self-simulation hypothesis, a simple form of implementation of the principle of efficient language is discussed in a self-referential geometric quasicrystalline state sum model in
Geometric State Sum Models from Quasicrystals
In light of the self-simulation hypothesis, a simple form implementation of the principle of efficient language is discussed in a self-referential geometric quasicrystalline state sum model in three
Approach to the cellular automaton interpretation using deep learning
  • H. Go
  • Computer Science
  • 2020
This paper shows that it is possible to reach covariant CNNs through the proposed method without implementing the specific algorithm, and demonstrates to learning process of cellular automata (CA) that could satisfy locality, time-reversibility through CNNs.


Characteristic forms and geometric invariants
Learning Structural Node Embeddings via Diffusion Wavelets
GraphWave is developed, a method that represents each node's network neighborhood via a low-dimensional embedding by leveraging heat wavelet diffusion patterns and mathematically proves that nodes with similar network neighborhoods will have similar GraphWave embeddings even though these nodes may reside in very different parts of the network, and the method scales linearly with the number of edges.
Learning the Einstein-Podolsky-Rosen correlations on a Restricted Boltzmann Machine
We construct a hidden variable model for the EPR correlations using a Restricted Boltzmann Machine. The model reproduces the expected correlations and thus violates the Bell inequality, as required
Precedence and Freedom in Quantum Physics
A new interpretation of quantum mechanics is proposed according to which precedence, freedom and novelty play central roles. This is based on a modification of the postulates for quantum theory given
On the separation of Einsteinian substructures
Within the spinorial version of the Cartan structure formulas with the built‐in (complex) Einstein vacuum equations some closed semi‐Einsteinian substructures are isolated and discussed. Then the
Mutual information, neural networks and the renormalization group
This work demonstrates a machine-learning algorithm capable of identifying the relevant degrees of freedom of a system and executing RG steps iteratively without any prior knowledge about the system, and applies the algorithm to classical statistical physics problems in one and two dimensions.
Extremal variety as the foundation of a cosmological quantum theory
Dynamical systems of a new kind are described, which are motivated by the problem of constructing diffeomorphism invariant quantum theories. These are based on the extremization of a non-local and
Unification of the State with the Dynamical Law
We address the question of why particular laws were selected for the universe, by proposing a mechanism for laws to evolve. Normally in physical theories, timeless laws act on time-evolving states.
Unification of gravity, gauge fields, and Higgs bosons
We consider a diffeomorphism invariant theory of a gauge field valued in a Lie algebra that breaks spontaneously to the direct sum of the spacetime Lorentz algebra, a Yang-Mills algebra, and their