Entangled Neural Networks from Multi-fold Universes to Biology
Stéphane Maes
Abstract
In a multi-fold universe, gravity emerges from Entanglement through the multi-fold mechanisms. As a result, gravity-like effects appear in between entangled particles, that they be real or virtual. Long range, massless gravity results from entanglement of massless virtual particles. Entanglement of massive virtual particles leads to massive gravity contributions at very smalls scales. Multi-folds mechanisms also result into a spacetime that is discrete, with a random walk fractal structure and non-commutative geometry that is Lorentz invariant and where spacetime nodes and particles can be modeled with microscopic black holes. All these recover General relativity at large scales, and semi-classical model remain valid till smaller scale than usually expected. Gravity can therefore be added to the Standard Model. This can contribute to resolving several open issues with the Standard Model without new Physics other than gravity. These considerations hints at a even stronger relationship between gravity and the Standard Model.Recently a controversial series of papers ended up proposing the possibility that the universe be a neural network. It is the result of observing that with an irreversible thermodynamics model of the learning process of the neural network (NN), it might appear possible to model quantum and classical physics, to observe the emergence of a General Relativistic spacetime with gravity, and plausibly to construct a generalized holographic principle beyond the AdS/CFT correspondence conjecture. The approach has been received with some skepticism.In this paper, we revisit the notion of NN in relationship to multi-fold universes, and illustrate how the multi-fold mechanism can be implemented with grafted NN. Relying on progress in biology and medicine, we argue that not only just NN can emulate NN universe, but also that it can provide new tools for AI, and new approaches to NNs, shallow or deep. It validates our multi-fold models and offer models for biological neurological models.