Explorations in High Energy Theory and Machine Learning

Northeastern University, Boston MA

Investigators

Abstract

This award funds the research activities of Professor Fabian Ruehle at Northeastern University. String theory has evolved into one of the most complex theories hitherto devised by mankind. It is the leading candidate to describe our Universe, from the smallest to the largest scales. Perhaps the most striking prediction of string theory is the existence of extra dimensions beyond length, width, and height. Owing to experimental results from particle colliders and cosmological observations, we know that these extra dimensions have to be small. Nevertheless, they encode all observable physics. Despite tremendous progress over the years, several questions remain unanswered. What is the geometry of these extra dimensions? Can their properties be used to rule out theories that seem consistent from our traditional three-dimensional point of view? What do these extra dimensions teach us about the most extreme environments in our Universe, such as near Black Holes? In answering these questions, scientists face two obstacles. First, advanced mathematical tools are needed in order to decode the implications of string theory. Second, even with these tools, some of the resulting equations are too complicated to solve with conventional techniques. As part of his research, Professor Ruehle will tackle these long-standing open questions by capitalizing on recent advances within search and optimization strategies in artificial intelligence and machine learning, paired with theoretical advances in the field. This project is also envisioned to have extensive broader impacts. Professor Ruehle will involve graduate students in his research, and in this way train the next generation of junior researchers in these important fields. The benefits and impact of recent advances of machine learning and artificial intelligence on both our everyday lives as well as in scientific research and discovery is remarkable and therefore in the national interest. Such research also opens up new avenues for inter- and multi-disciplinary research and collaborations. Indeed, such research has not only produced techniques for classifying the properties of extra dimensions but has even led to techniques for identifying biomarkers for severe disease progression of COVID-19. More technically, Professor Ruehle will use neural networks to describe metrics of compactification manifolds with reduced holonomy or special structure. These metrics are needed to solve the equations of motion of string theory. They enter in observable data (such as Yukawa couplings and massive string excitations), but also play an important role in more mathematical questions (for example mirror symmetry, or non-calibrated cycles), which Ruehle will explore. Moreover, Ruehle will develop and adapt search algorithms from data science, AI, and quantum computing to derive intelligent search strategies for combinatorial problems arising in string theory that cannot be tackled with conventional techniques. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.