City Research Online

Abstract

This thesis explores two distinct but powerful approaches used in theoretical physics and mathematics to study complex systems: AdS/CFT integrability and Machine Learning. The first part starts with an introduction to AdS/CFT integrability, reviewing the integrability
in AdS5 × S5 and AdS3 × S3 × T4 backgrounds, which is then used to present the novel kinematical structure of the mixed-flux AdS3 × S3 × T4 background and a previously unknown solution for the odd part of the dressing factor of the exact worldsheet S-matrix. The second part applies Machine Learning to two combinatorial problems in Clifford
algebras and error-correcting codes. After a brief introduction of Machine Learning methods, we explore the potential of network classification and principal component analysis in the study of Clifford geometric invariants of Coxeter elements in the root systems of ADE
algebras, demonstrating the viability of the computational approach and the potential for further analytical exploration. We then address the problem of searching for new champion codes of generalised toric codes. We develop a pipeline that employs a Transformer Deep
Learning architecture and Genetic Algorithm to find new champion codes for F8 generalised toric codes.

Publication Type:	Thesis (Doctoral)
Departments:	School of Science & Technology > Department of Mathematics School of Science & Technology > School of Science & Technology Doctoral Theses Doctoral Theses

Export

Downloads