M-Theory Research

This directory contains TensorFlow-based code to address some research problems in M-Theory / Superstring Theory / Supergravity / Quantum Gravity.

Broadly speaking, M-Theory is all about the very rich mathematical structure that arises if one tries to reconcile the physical principles of Quantum Mechanics, General Relativity, and Supersymmetry. In terms of "inputs" and "data", this research should be regarded as Pure Mathematics, i.e. there is no dependency on "measurement" (or even "user") data.

Still, despite this research not using any data examples (for learning or otherwise), Google's TensorFlow Machine Learning technology is sufficiently generic to be a very useful tool to address some research questions in this domain that can/should be studied numerically.

Downloading and Installation

A simple (albeit somewhat strange) way to download only this part of the google-research github repository is:

svn export https://github.com/google-research/google-research/trunk/m_theory

Then, the Python environment can be set up as follows:

virtualenv -p python3 env
source env/bin/activate

pip3 install -r m_theory_lib/requirements.txt

Structure

• dim4/so8_supergravity_extrema/

  Code for the scalar potential of the de Wit - Nicolai model, SO(8)-gauged N=8 Supergravity in 3+1-dimensional spacetime.

  Article: "SO(8) Supergravity and the Magic of Machine Learning" (https://arxiv.org/abs/1906.00207).

  Demo: This will run a small demo search for a few solutions, plus analysis of one of those obtained. Output (providing location data and particle properties) will be in the directory EXAMPLE_SOLUTIONS.

  python3 -m dim4.so8_supergravity_extrema.code.extrema

• wrapped_branes/

  Code for analyzing the potentials of the models constructed in https://arxiv.org/abs/1906.08900 and https://arxiv.org/abs/1009.3805 by wrapping M5-branes.

  Run via: python3 -i -m wrapped_branes.wrapped_branes {problem_name} with {problem_name} one of: dim7, cgr-S2, cgr-R2, cgr-H2, dgkv-S3, dgkv-R3, dgkv-H3.