European Commission logo
English English
CORDIS - EU research results

Machine Learning for String Field Theory and for the String- and F-Theory landscapes

Project description

Artificial intelligence could help reduce the computational complexity of string theory

String theory is one of the most promising and controversial ideas in modern physics, attempting to unify quantum gravity and all different forces, particles and interactions under the same framework. The theoretical extra dimensions the theory proposes are hard to visualise, and there are many possible ways these various geometries could be folded in on themselves. String field theory is the newest approach to describing string theory. The EU-funded ML4SFT project will tap into the potential of artificial intelligence to help scientists overcome computational hurdles in string theory. In particular, it will elaborate tools to construct the string field theory action while also enhancing understanding of closed string field theory.


"String theory is one of the leading endeavours in theoretical physics to construct a theory of quantum gravity unified with all interactions and matter. As such, it provides a complete description of the Universe and of its content. However, while all ingredients are present, the details for a precise contact with the Standard Model and with our Universe are missing, mostly because the number of possible realizations is huge and no selection mechanism is known. Moreover, progress is further hindered, for the description of string theory as a field theory - arguably its most fundamental formulation - is very intricate. Since the most immediate difficulties are computational - it boils down to studying statistics of geometries (the ""string landscape"") and to approximating functions and geometries (to build a string field theory) - machine learning seems to provide an adequate framework to address the challenges faced by string theory.
The first aspect of this project is to elaborate machine learning tools for constructing the string field theory action while deepening in parallel our analytic understanding of closed string field theory. For the latter, the main objective is to include auxiliary fields and to investigate whether the action can be made cubic. The second aspect is to design machine learning algorithms to map the string landscape.
This project holds the promise of important developments in our understanding of string theory and in its applications to phenomenology. It is located at the intersection of multiple disciplines - theoretical physics, mathematics (Riemann surfaces and homotopy algebras) and machine learning. The choice of institutions and supervisors reflect this interdisciplinary aspect: Prof. Zwiebach is an expert in string (field) theory and in the moduli space geometry, while Dr. Tamaazousti is a specialist of machine learning. Furthermore, both institutions are renowned in these domains."


Net EU contribution
€ 275 619,84
75015 PARIS 15

See on map

Ile-de-France Ile-de-France Paris
Activity type
Research Organisations
Total cost
€ 275 619,84

Partners (1)